We introduce SEAS, using ShapE-Aligned Supervision, to enhance appearance-based person re-identification. When recognizing an individual's identity, existing methods primarily rely on appearance, which can be influenced by the background environment due to a lack of body shape awareness. Although some methods attempt to incorporate other modalities, such as gait or body shape, they encode the additional modality separately, resulting in extra computational costs and lacking an inherent connection with appearance. In this paper, we explore the use of implicit 3-D body shape representations as pixel-level guidance to augment the extraction of identity features with body shape knowledge, in addition to appearance. Using body shape as supervision, rather than as input, provides shape-aware enhancements without any increase in computational cost and delivers coherent integration with pixel-wise appearance features. Moreover, for video-based person re-identification, we align pixel-level features across frames with shape awareness to ensure temporal consistency. Our results demonstrate that incorporating body shape as pixel-level supervision reduces rank-1 errors by 32.8% for frame-based and by 27.2% for video-based re-identification tasks, respectively, and can also be generalized to other existing appearance-based person re-identification methods.

Face Anti-Spoofing (FAS) is pivotal in safeguarding facial recognition systems against presentation attacks. While domain generalization (DG) methods have been developed to enhance FAS performance, they predominantly focus on learning domain-invariant features during training, which may not guarantee generalizability to unseen data that differs largely from the source distributions. Our insight is that testing data can serve as a valuable resource to enhance the generalizability beyond mere evaluation for DG FAS. In this paper, we introduce a novel Test-Time Domain Generalization (TTDG) framework for FAS, which leverages the testing data to boost the model's generalizability. Our method, consisting of Test-Time Style Projection (TTSP) and Diverse Style Shifts Simulation (DSSS), effectively projects the unseen data to the seen domain space. In particular, we first introduce the innovative TTSP to project the styles of the arbitrarily unseen samples of the testing distribution to the known source space of the training distributions. We then design the efficient DSSS to synthesize diverse style shifts via learnable style bases with two specifically designed losses in a hyperspherical feature space. Our method eliminates the need for model updates at the test time and can be seamlessly integrated into not only the CNN but also ViT backbones. Comprehensive experiments on widely used cross-domain FAS benchmarks demonstrate our method's state-of-the-art performance and effectiveness.

Recent advancements in domain generalization (DG) for face anti-spoofing (FAS) have garnered considerable attention. Traditional methods have focused on designing learning objectives and additional modules to isolate domain-specific features while retaining domain-invariant characteristics in their representations. However, such approaches often lack guarantees of consistent maintenance of domain-invariant features or the complete removal of domain-specific features. Furthermore, most prior works do not ensure convergence to a local flat minimum, which has been shown to be advantageous for DG. In this paper, we introduce GAC-FAS, a novel learning objective that encourages the model to converge towards an optimal flat minimum without necessitating additional learning modules. Unlike conventional sharpness-aware minimizers, GAC-FAS identifies ascending points for each domain and regulates the generalization gradient updates at these points to align coherently with empirical risk minimization (ERM) gradient updates. This unique approach specifically guides the model to be robust against domain shifts. We demonstrate the efficacy of GAC-FAS through rigorous testing on challenging cross-domain FAS datasets, where it establishes SoTA performance on FAS task.

Gait recognition stands as one of the most pivotal remote identification technologies and progressively expands across research and industrial communities. However, existing gait recognition methods heavily rely on task-specific upstream driven by supervised learning to provide explicit gait representations, which inevitably introduce expensive annotation costs and potentially cause cumulative errors. Escaping from this trend, this work explores effective gait representations based on the all-purpose knowledge produced by task-agnostic Large Vision Models (LVMs) and proposes a simple yet efficient gait framework, termed \textbf{BigGait}. Specifically, the Gait Representation Extractor (GRE) in BigGait effectively transforms all-purpose knowledge into implicit gait features in an unsupervised manner, drawing from design principles of established gait representation construction approaches. Experimental results on CCPG, CAISA-B* and SUSTech1K indicate that BigGait significantly outperforms the previous methods in both self-domain and cross-domain tasks in most cases, and provides a more practical paradigm for learning the next-generation gait representation. Eventually, we delve into prospective challenges and promising directions in LVMs-based gait recognition, aiming to inspire future work in this emerging topic. All the source code will be available.

Face Anti-Spoofing (FAS) is crucial for securing face recognition systems against presentation attacks. With advancements in sensor manufacture and multi-modal learning techniques, many multi-modal FAS approaches have emerged. However, they face challenges in generalizing to unseen attacks and deployment conditions. These challenges arise from (1) modality unreliability, where some modality sensors like depth and infrared undergo significant domain shifts in varying environments, leading to the spread of unreliable information during cross-modal feature fusion, and (2) modality imbalance, where training overly relies on a dominant modality hinders the convergence of others, reducing effectiveness against attack types that are indistinguishable by sorely using the dominant modality.To address modality unreliability, we propose the Uncertainty-Guided Cross-Adapter (U-Adapter) to recognize unreliably detected regions within each modality and suppress the impact of unreliable regions on other modalities. For modality imbalance, we propose a Rebalanced Modality Gradient Modulation (ReGrad) strategy to rebalance the convergence speed of all modalities by adaptively adjusting their gradients.Besides, we provide the first large-scale benchmark for evaluating multi-modal FAS performance under domain generalization scenarios. Extensive experiments demonstrate that our method outperforms state-of-the-art methods. Source codes and protocols are released on https://github.com/OMGGGGG/mmdg.

Domain generalization (DG) based Face Anti-Spoofing (FAS) aims to improve the model's performance on unseen domains. Existing methods either rely on domain labels to align domain-invariant feature spaces, or disentangle generalizable features from the whole sample, which inevitably lead to the distortion of semantic feature structures and achieve limited generalization. Instead of directly manipulating visual features, we make use of large-scale vision-language models (VLMs) like CLIP and leverage the textual feature to dynamically adjust the classifier's weights for exploring generalizable visual features. Specifically, we propose a novel Class Free Prompt Learning (CFPL) paradigm for DG FAS, which utilizes two lightweight transformers, namely Content Q-Former (CQF) and Style Q-Former (SQF), to learn the different semantic prompts conditioned on content and style features by using a set of learnable query vectors, respectively. Thus, the generalizable prompt can be learned by two improvements: (1) A Prompt-Text Matched (PTM) supervision is introduced to ensure CQF learns visual representation that is most informative of the content description. (2) A Diversified Style Prompt (DSP) technology is proposed to diversify the learning of style prompts by mixing feature statistics between instance-specific styles. Finally, the learned text features modulate visual features to generalization through the designed Prompt Modulation (PM). Extensive experiments show that the CFPL is effective and outperforms the state-of-the-art methods on several cross-domain datasets.

Reconstructing the viewed images from human brain activity bridges human and computer vision through the Brain-Computer Interface. The inherent variability in brain function between individuals leads existing literature to focus on acquiring separate models for each individual using their respective brain signal data, ignoring commonalities between these data. In this article, we devise Psychometry, an omnifit model for reconstructing images from functional Magnetic Resonance Imaging (fMRI) obtained from different subjects. Psychometry incorporates an omni mixture-of-experts (Omni MoE) module where all the experts work together to capture the inter-subject commonalities, while each expert associated with subject-specific parameters copes with the individual differences. Moreover, Psychometry is equipped with a retrieval-enhanced inference strategy, termed Ecphory, which aims to enhance the learned fMRI representation via retrieving from prestored subject-specific memories. These designs collectively render Psychometry omnifit and efficient, enabling it to capture both inter-subject commonality and individual specificity across subjects. As a result, the enhanced fMRI representations serve as conditional signals to guide a generation model to reconstruct high-quality and realistic images, establishing Psychometry as state-of-the-art in terms of both high-level and low-level metrics. Our code will be released.

In this paper, we address the challenge of making ViT models more robust to unseen affine transformations. Such robustness becomes useful in various recognition tasks such as face recognition when image alignment failures occur. We propose a novel method called KP-RPE, which leverages key points (e.g.~facial landmarks) to make ViT more resilient to scale, translation, and pose variations. We begin with the observation that Relative Position Encoding (RPE) is a good way to bring affine transform generalization to ViTs. RPE, however, can only inject the model with prior knowledge that nearby pixels are more important than far pixels. Keypoint RPE (KP-RPE) is an extension of this principle, where the significance of pixels is not solely dictated by their proximity but also by their relative positions to specific keypoints within the image. By anchoring the significance of pixels around keypoints, the model can more effectively retain spatial relationships, even when those relationships are disrupted by affine transformations. We show the merit of KP-RPE in face and gait recognition. The experimental results demonstrate the effectiveness in improving face recognition performance from low-quality images, particularly where alignment is prone to failure.Code and pre-trained models will be shared.

Person Re-Identification (ReID) holds critical importance in computer vision with pivotal applications in public safety and crime prevention. Traditional ReID methods, reliant on appearance attributes such as clothing and color, encounter limitations in long-term scenarios and dynamic environments. To address these challenges, we propose CLIP3DReID, an innovative approach that enhances person ReID by integrating linguistic descriptions with visual perception, leveraging pretrained CLIP model for knowledge distillation. Our method first employs CLIP to automatically label body shapes with linguistic descriptors. We then apply optimal transport theory to align the student model's local visual features with shape-aware tokens derived from CLIP's linguistic output. Additionally, we align the student model's global visual features with those from the CLIP image encoder and the 3D SMPL identity space, fostering enhanced domain robustness. CLIP3DReID notably excels in discerning discriminative body shape features, achieving state-of-the-art results in person ReID. Our approach represents a significant advancement in ReID, offering robust solutions to existing challenges and setting new directions for future research. The code and models will be released upon publication.

Periocular and face are complementary biometrics for identity management, albeit with inherent limitations, notably in scenarios involving occlusion due to sunglasses or masks. In response to these challenges, we introduce Flexible Biometric Recognition (FBR), a novel framework designed to advance conventional face, periocular, and multimodal face-periocular biometrics across both intra- and cross-modality recognition tasks. FBR strategically utilizes the Multimodal Fusion Attention (MFA) and Multimodal Prompt Tuning (MPT) mechanisms within the Vision Transformer architecture. MFA facilitates the fusion of modalities, ensuring cohesive alignment between facial and periocular embeddings while incorporating soft-biometrics to enhance the model's ability to discriminate between individuals. The fusion of three modalities is pivotal in exploring interrelationships between different modalities. Additionally, MPT serves as a unifying bridge, intertwining inputs and promoting cross-modality interactions while preserving their distinctive characteristics. The collaborative synergy of MFA and MPT enhances the shared features of the face and periocular, with a specific emphasis on the ocular region, yielding exceptional performance in both intra- and cross-modality recognition tasks. Rigorous experimentation across four benchmark datasets validates the noteworthy performance of the FBR model. The source code is available at https://github.com/MIS-DevWorks/FBR.

Face anti-spoofing (FAS) is critical to applications which heavily rely on the authenticity of detected faces. Many FAS methods have focused on learning discriminative features from both live and spoof training data. However, since not every possible attack type is available in the training stage, these FAS methods usually fail to detect unseen attacks in the inference stage. In comparison, One-Class Classification, where the training data are from only a single positive class (e.g., live faces), enables a more practical setting for FAS to detect whether a test face image belongs to the live class or not. In this paper, we address the one-class FAS detection problem and propose a novel One-Class Spoof Cue Map estimation Network (OC-SCMNet) to detect various spoof attacks by learning exclusively from the live class. We construct OC-SCMNet with one latent feature extractor, one Spoof Cue Map (SCM) estimator, and one SCM-guided generator. Our first goal is to learn to extract latent spoof features from live images so that their estimated SCMs should have zero responses. To avoid trapping to a trivial solution, we devise a novel SCM-guided feature learning by combining many SCMs as pseudo ground-truths to guide a conditional generator to generate non-trivial latent spoof features for spoof data. Our second goal is to approximately simulate the potential out-of-distribution spoof attacks under one-class constraint. To this end, we propose using a memory bank to dynamically preserve a set of sufficiently “independent” latent spoof features to encourage the generator to probe the latent spoof feature space. Extensive experiments conducted on eight FAS benchmark datasets demonstrate that the proposed OC-SCMNet not only outperforms previous one-class FAS methods but also achieves comparable performances to state-of-the-art two-class FAS method.

In this work, we study a novel problem which focuses on person identification while performing daily activities. Learning biometric features from RGB videos is challenging due to spatio-temporal complexity and presence of appearance biases such as clothing color and background. We propose ABNet, a novel framework which leverages disentanglement of biometric and non-biometric features to perform effective person identification from daily activities. ABNet relies on a bias-less teacher to learn biometric features from RGB videos and explicitly disentangle nonbiometric features with the help of biometric distortion. In addition, ABNet also exploits activity prior for biometrics which is enabled by joint biometric and activity learning. We perform comprehensive evaluation of the proposed approach across five different datasets which are derived from existing activity recognition benchmarks. Furthermore, we extensively compare ABNet with existing works in person identification and demonstrate its effectiveness for activity-based biometrics across all five datasets.

The widespread adoption of face recognition has led to increasing privacy concerns, as unauthorized access to face images can expose sensitive personal information. This paper explores face image protection against viewing and recovery attacks. Inspired by image compression, we propose creating a visually uninformative face image through feature subtraction between an original face and its model-produced regeneration. Recognizable identity features within the image are encouraged by co-training a recognition model on its high-dimensional feature representation. To enhance privacy, the high-dimensional representation is crafted through random channel shuffling, resulting in randomized recognizable images devoid of attacker-leverageable texture details. We distill our methodologies into a novel privacy-preserving face recognition method, MinusFace. Extensive experiments demonstrate its high recognition accuracy and effective privacy protection.

The premise for the great advancement of molecular machine learning is dependent on a considerable amount of labeled data. In many real-world scenarios, the labeled molecules are limited in quantity or laborious to derive. Recent pseudo-labeling methods are usually designed based on a single domain knowledge, thereby failing to understand the comprehensive molecular configurations and limiting their adaptability to generalize across diverse biochemical context. To this end, we introduce an innovative paradigm for dealing with the molecule pseudo-labeling, named as Molecular Data Programming (MDP). In particular, we adopt systematic supervision sources via crafting multiple graph labeling functions, which covers various molecular structural knowledge of graph kernels, molecular fingerprints, and topological features. Each of them creates an uncertain and biased labels for the unlabeled molecules. To address the decision conflicts among the diverse pseudo-labels, we design a label synchronizer to differentiably model confidences and correlations between the labeling functions, which yields probabilistic molecular labels to adapt for specific applications. These probabilistic molecular labels are used to train a molecular classifier for improving its generalization capability. On eight benchmark datasets, we empirically demonstrate the effectiveness of MDP on the weakly supervised molecule classification tasks, achieving an average improvement of $9.5\%$.

Protein representation learning is a challenging task that aims to capture the structure and function of proteins from their amino acid sequences. Previous methods largely ignored the fact that not all amino acids are equally important for protein folding and activity. In this article, we propose a neural clustering framework that can automatically discover the critical components of a protein by considering both its primary and tertiary structure information. Our framework treats a protein as a graph, where each node represents an amino acid and each edge represents a spatial or sequential connection between amino acids. We then apply an iterative clustering strategy to group the nodes into clusters based on their 1D and 3D positions and assign scores to each cluster. We select the highest-scoring clusters and use their medoid nodes for the next iteration of clustering, until we obtain a hierarchical and informative representation of the protein. We evaluate our framework on four protein-related tasks: protein fold classification, enzyme reaction classification, gene ontology term prediction, and enzyme commission number prediction. Experimental results demonstrate that our method achieves state-of-the-art performance. Our code will be released.

Principal component analysis (PCA), along with its extensions to manifolds and outlier contaminated data, have been indispensable in computer vision and machine learning. In this work, we present a unifying formalism for PCA and its variants, and introduce a framework based on the flags of linear subspaces, i.e., a hierarchy of nested linear subspaces of increasing dimension, which not only allows for a common implementation but also yields novel variants, not explored previously. We begin by generalizing traditional PCA methods that either maximize variance or minimize reconstruction error. We expand these interpretations to develop a wide array of new dimensionality reduction algorithms by accounting for outliers and the data manifold. To devise a common computational approach, we recast robust and dual forms of PCA as optimization problems on flag manifolds. We then integrate tangent space approximations of principal geodesic analysis (tangent-PCA) into this flag-based framework, creating novel robust and dual geodesic PCA variations. The remarkable flexibility offered by the `flagification' introduced here enables even more algorithmic variants identified by specific flag types. Last but not least, we propose an effective convergent solver for these flag-formulations employing the Stiefel manifold. Our empirical results on both real-world and synthetic scenarios, demonstrate the superiority of our novel algorithms, especially in terms of robustness to outliers on manifolds.

Recently, convolutional neural networks (CNNs) with large size kernels have attracted much attention in the computer vision field, following the success of the Vision Transformers. Large kernel CNNs have been reported to perform well in downstream vision tasks as well as in classification performance. The reason for the high-performance of large kernel CNNs in downstream tasks has been attributed to the large effective receptive field (ERF) produced by large size kernels, but this view has not been fully tested. We therefore revisit the performance of large kernel CNNs in downstream task, focusing on the weakly supervised object localization (WSOL) task. WSOL, a difficult downstream task that is not fully supervised, provides a new angle to explore the capabilities of the large kernel CNNs. Our study compares the modern large kernel CNNs ConvNeXt, RepLKNet, and SLaK to test the validity of the naive expectation that ERF size is important for improving downstream task performance. Our analysis of the factors contributing to high performance provides a different perspective, in which the main factor is feature map improvement. Furthermore, we find that modern CNNs are robust to the CAM problems of local regions of objects being activated, which has long been discussed in WSOL. CAM is the most classic WSOL method, but because of the above-mentioned problems, it is often used as a baseline method for comparison. However, experiments on the CUB-200-2011 dataset show that simply combining a large kernel CNN, CAM, and simple data augmentation methods can achieve performance (90.99\% MaxBoxAcc) comparable to the latest WSOL method, which is CNN-based and requires special training or complex post-processing.

Addressing pose ambiguity in 6D object pose estimation from single RGB images presents a significant challenge, particularly due to object symmetries or occlusions. In response, we introduce a novel score-based diffusion method applied to the $SE(3)$ group, marking the first application of diffusion models to $SE(3)$ within the image domain, specifically tailored for pose estimation tasks. Extensive evaluations demonstrate the method's efficacy in handling pose ambiguity, mitigating perspective-induced ambiguity, and showcasing the robustness of our surrogate Stein score formulation on $SE(3)$. This formulation not only improves the convergence of denoising process but also enhances computational efficiency. Thus, we pioneer a promising strategy for 6D object pose estimation.

The goal of multi-task learning is to learn diverse tasks within a single unified network. As each task has its own unique objective function, conflicts emerge during training, resulting in negative transfer among them. Earlier research identified these conflicting gradients in shared parameters between tasks and attempted to realign them in the same direction. However, we prove that such optimization strategies lead to sub-optimal Pareto solutions due to their inability to accurately determine the individual contributions of each parameter across various tasks. In this paper, we propose the concept of task priority to evaluate parameter contributions across different tasks. To learn task priority, we identify the type of connections related to links between parameters influenced by task-specific losses during backpropagation. The strength of connections is gauged by the magnitude of parameters to determine task priority. Based on these, we present a new method named connection strength-based optimization for multi-task learning which consists of two phases. The first phase learns the task priority within the network, while the second phase modifies the gradients while upholding this priority. This ultimately leads to finding new Pareto optimal solutions for multiple tasks. Through extensive experiments, we show that our approach greatly enhances multi-task performance in comparison to earlier gradient manipulation methods.

In the literature, points and conics have been major features for camera geometric calibration. Although conics are more informative features than points, the loss of the conic property under distortion has critically limited the utility of conic features in camera calibration. Many existing approaches addressed conic-based calibration by ignoring distortion or introducing 3D spherical targets to circumvent this limitation. In this paper, we present a novel formulation for conic-based calibration using moments. Our derivation is based on the mathematical finding that the first moment can be estimated without bias even under distortion. This allows us to track moment changes during projection and distortion, ensuring the preservation of the first moment of the distorted conic. With an unbiased estimator, the circular patterns can be accurately detected at the sub-pixel level and can now be fully exploited for an entire calibration pipeline, resulting in significantly improved calibration. The entire code is readily available from https://github.com/DisCoCal/DisCoCal.

Low-light scenes are prevalent in the real world applications (e.g., autonomous driving, security cameras at night). Recently, multi-object tracking in various practical use cases have garnered a lot of attention, but the multi-object tracking in the dark scenes is rarely considered. In this paper, we focus on multi-object tracking in the dark scenes. To address the lack of dataset, we build a \textbf{L}ow-light \textbf{M}ulti-\textbf{O}bject \textbf{T}racking (\textbf{LMOT}) dataset. LMOT provides highly aligned low-light video pairs captured by our dual-camera system and high quality multi-object tracking annotations for all videos. Then, we propose a Low-light Multi-object Tracking method, termed as \textbf{LTrack}. We introduce the adaptive low-pass downsample module to enhance low frequency components of images outside the sensor noises. Degradation suppression learning strategy enable the model to learn invariant information under the noise disturbance and image quality degradation. These components improve the robustness of multi-object tracking in the dark scenes. We conduct a comprehensive analysis to our LMOT dataset and proposed LTrack. Experimental results demonstrate the superiority of the proposed method and its competitiveness in real night low-light scenes.

Visible-Infrared Person Re-identification (VI-ReID) is a challenging cross-modal pedestrian retrieval task, due to significant intra-class variations and cross-modal discrepancies among different cameras. Existing works mainly focus on embedding images of different modalities into a unified space to mine modality-shared features. They only seek distinctive information within these shared features, while ignoring the identity-aware useful information that is implicit in the modality-specific features. To address this issue, we propose a novel Implicit Discriminative Knowledge Learning (IDKL) network to uncover and leverage the implicit discriminative information contained within the modality-specific. First, we extract modality-specific and modality-shared features using a novel dual-stream network. Then, the modality-specific features undergo purification to reduce their modality style discrepancies while preserving identity-aware discriminative knowledge. Subsequently, this kind of implicit knowledge is distilled into the modality-shared feature to enhance its distinctiveness. Finally, an alignment loss is proposed to minimize modality discrepancy on modality-shared features. Extensive experiments on multiple public datasets demonstrate the superiority of IDKL network over the state-of-the-art methods.

Estimating the relative camera pose from $n \geq 5$ correspondences between two calibrated views is a fundamental task in computer vision. This process typically involves two stages: 1) estimating the essential matrix between the views, and 2) disambiguating among the four candidate relative poses that satisfy the epipolar geometry. In this paper, we demonstrate a novel approach that, for the first time, bypasses the second stage. Specifically, we show that it is possible to directly estimate the correct relative camera pose from correspondences without needing a post-processing step to enforce the cheirality constraint on the correspondences.Building on recent advances in certifiable non-minimal optimization, we frame the relative pose estimation as a Quadratically Constrained Quadratic Program (QCQP). By applying the appropriate constraints, we ensure the estimation of a camera pose that corresponds to a valid 3D geometry and that is globally optimal when certified. We validate our method through exhaustive synthetic and real-world experiments, confirming the efficacy, efficiency and accuracy of the proposed approach. Our code can be found in the supp. material and will be released.

In the realm of computer vision, Neural Fields have gained prominence as a contemporary tool harnessing neural networks for signal representation. Despite the remarkable progress in adapting these networks to solve a variety of problems, the field still lacks a comprehensive theoretical framework. This article aims to address this gap by delving into the intricate interplay between initialization and activation, providing a foundational basis for the robust optimization of Neural Fields. Our theoretical insights reveal a deep-seated connection among network initialization, architectural choices, and the optimization process, emphasizing the need for a holistic approach when designing cutting-edge Neural Fields.

Image and video analysis requires not only accurate object detection but also the understanding of relationships among detected objects.Common solutions to relation modeling typically resort to stand-alone object detectors followed by non-differentiable post-processing techniques. Recently introduced detection transformers (DETR) perform end-to-end object detection based on a bipartite matching loss. Using traditional object detection methods or even DETR-based models (deformable detr, dino, etc.) lacks the ability to detect objects and their relationships directly.In this paper, we build on this approach and extend it to the joint detection of objects and their relationships.A naive extension of DETR to object relations, however, leads to a NP-hard problem. To this end, we propose an approximate solution based on bipartite matching.While our method can generalize to an arbitrary number of objects, we here focus on the modeling of object pairs and their relations.In particular, we apply our method PairDETR to the problem of detecting human bodies, faces, and associations between bodies and faces of the same person. Our approach not only eliminates the need for hand-designed post-processing but also achieves excellent results for body-face associations.We evaluate PairDETR on the challenging CrowdHuman and CityPersons datasets and demonstrate a significant improvement over the state of the art. Our training code and pre-trained models will become publicly available.

Despite significant advancements in text-to-motion synthesis, generating language-guided human motion within 3D environments poses substantial challenges. These challenges stem primarily from (i) the absence of powerful generative models capable of jointly modeling natural language, 3D scenes, and human motion, and (ii) the generative models' intensive data requirements contrasted with the scarcity of comprehensive, high-quality, language-scene-motion datasets. To tackle these issues, we introduce a novel two-stage framework that employs scene affordance as an intermediate representation, effectively linking 3D scene grounding and conditional motion generation. Our framework comprises an Affordance Diffusion Model (ADM) for predicting explicit affordance map and an Affordance-to-Motion Diffusion Model (AMDM) for generating plausible human motions. By leveraging scene affordance maps, our method overcomes the difficulty in generating human motion under multimodal condition signals, especially when training with limited data lacking extensive language-scene-motion pairs. Our extensive experiments demonstrate that our approach consistently outperforms all baselines on established benchmarks, including HumanML3D and HUMANISE. Additionally, we validate our model's exceptional generalization capabilities on a specially curated evaluation set featuring previously unseen descriptions and scenes.

We present OAKINK2, a dataset of bimanual object manipulation tasks for complex daily activities. In pursuit of constructing the complex tasks into a structured representation, OAKINK2 introduces three level of abstraction to organize the manipulation tasks: Affordance, Primitive Task, and Complex Task. OAKINK2 features on an object-centric perspective for decoding the complex tasks, treating them as a sequence of object affordance fulfillment. The first level, Affordance, outlines the functionalities that objects in the scene can afford, the second level, Primitive Task, describes the minimal interaction units that humans interact with the object to achieve its affordance, and the third level, Complex Task, illustrates how Primitive Tasks are composed and interdependent. OAKINK2 dataset provides multi-view image streams and precise pose annotations for the human body, hands and various interacting objects. This extensive collection supports applications such as interaction reconstruction and motion synthesis. Based on the 3-level abstraction of OAKINK2, we explore a task-oriented framework for Complex Task Completion (CTC). CTC aims to generate a sequence of bimanual manipulation to achieve task objectives. Within the CTC framework, we employ Large Language Models (LLMs) to decompose the complex task objectives into sequences of Primitive Tasks and have developed a Motion Fulfillment Model that generates bimanual hand motion for each Primitive Task. OAKINK2 datasets and models are available at https://oakink.net/v2.

Conditional human motion generation is an important topic with many applications in virtual reality, gaming, and robotics. While prior works have focused on generating motion guided by text, music, or scenes, these typically result in isolated motions confined to short durations. Instead, we address the generation of long, continuous sequences that are guided by a series of varying conditions. In this context, we introduce FlowMDM, the first diffusion-based model that generates seamless Human Motion Compositions (HMC) without any postprocessing or redundant denoising steps. For this, we introduce the Blended Positional Encodings, a technique that leverages both absolute and relative positional encodings in the denoising chain. More specifically, global motion coherence is recovered at the absolute stage, whereas smooth and realistic transitions are built at the relative stage. As a result, we achieve state-of-the-art results in terms of accuracy, realism, and smoothness on the Babel and HumanML3D datasets. FlowMDM excels when trained with only a single condition per motion sequence thanks to its Pose-Centric Cross-ATtention, which makes it robust against varying conditions at inference time. Finally, to address the limitations of existing HMC metrics, we propose two new metrics: the Peak Jerk and the Area Under the Jerk, for detecting abrupt transitions.

Neural Radiance Fields (NeRFs) excel in photorealistically rendering static scenes. However, rendering dynamic, long-duration radiance fields on ubiquitous devices remains challenging, due to data storage and computational constraints. In this paper, we introduce VideoRF, the first approach to enable real-time streaming and rendering of dynamic human-centric radiance fields on mobile platforms. At the core is a serialized 2D feature image stream representing the 4D radiance field all in one. We introduce a tailored training scheme directly applied to this 2D domain to impose the temporal and spatial redundancy of the feature image stream. By leveraging the redundancy, we show that the feature image stream can be efficiently compressed by 2D video codecs, which allows us to exploit video hardware accelerators to achieve real-time decoding. On the other hand, based on the feature image stream, we propose a novel rendering pipeline for VideoRF, which has specialized space mappings to query radiance properties efficiently. Paired with a deferred shading model, VideoRF has the capability of real-time rendering on mobile devices thanks to its efficiency. We have developed a real-time interactive player that enables online streaming and rendering of dynamic scenes, offering a seamless and immersive free-viewpoint experience across a range of devices, from desktops to mobile phones. Our project page is available at https://aoliao12138.github.io/VideoRF/.

We have recently seen tremendous progress in realistic text-to-motion generation. Yet, the existing methods often fail or produce implausible motions with unseen text inputs, which limits the applications. In this paper, we present OMG, a novel framework, which enables compelling motion generation from zero-shot open-vocabulary text prompts. Our key idea is to carefully tailor the pretrain-then-finetune paradigm into the text-to-motion generation. At the pre-training stage, our model improves the generation ability by learning the rich out-of-domain inherent motion traits. To this end, we scale up a large unconditional diffusion model up to 1B parameters, so as to utilize the massive unlabeled motion data up to over 20M motion instances. At the subsequent fine-tuning stage, we introduce motion ControlNet, which incorporates text prompts as conditioning information, through a trainable copy of the pre-trained model and the proposed novel Mixture-of-Controllers (MoC) block. MoC block adaptively recognizes various ranges of the sub-motions with a cross-attention mechanism and processes them separately with the text-token-specific experts. Such a design effectively aligns the CLIP token embeddings of text prompts to various ranges of compact and expressive motion features. Extensive experiments demonstrate that our OMG achieves significant improvements over the state-of-the-art methods on zero-shot text-to-motion generation. Project page:https://tr3e.github.io/omg-page.

Since humans interact with diverse objects every day, the holistic 3D capture of these interactions is important to understand and model human behaviour. However, most existing methods for hand-object reconstruction from RGB either assume pre-scanned object templates or heavily rely on limited 3D hand-object data, restricting their ability to scale and generalize to more unconstrained interaction settings. To address this, we introduce HOLD -- the first category-agnostic method that reconstructs an articulated hand and an object jointly from a monocular interaction video. We develop a compositional articulated implicit model that can reconstruct disentangled 3D hands and objects from 2D images. We also further incorporate hand-object constraints to improve hand-object poses and consequently the reconstruction quality. Our method does not rely on any 3D hand-object annotations while significantly outperforming fully-supervised baselines in both in-the-lab and challenging in-the-wild settings. Moreover, we qualitatively show its robustness in reconstructing from in-the-wild videos. See https://github.com/zc-alexfan/hold for code, data, models, and updates.

Recent advances in neural rendering have improved both training and rendering times by orders of magnitude. While these methods demonstrate state-of-the-art quality and speed, they are designed for photogrammetry of static scenes and do not generalize well to freely moving humans in the environment. In this work, we introduce Human Gaussian Splats (HUGS) that represents an animatable human together with the scene using 3D Gaussian Splatting (3DGS). Our method takes only a monocular video with a small number of (50-100) frames, and it automatically learns to disentangle the static scene and a fully animatable human avatar within 30 minutes. We utilize the SMPL body model to initialize the human Gaussians. To capture details that are not modeled by SMPL (e.g., cloth, hairs), we allow the 3D Gaussians to deviate from the human body model. Utilizing 3D Gaussians for animated humans brings new challenges, including the artifacts created when articulating the Gaussians. We propose to jointly optimize the linear blend skinning weights to coordinate the movements of individual Gaussians during animation. Our approach enables novel-pose synthesis of human and novel view synthesis of both the human and the scene. We achieve state-of-the-art rendering quality with a rendering speed of 32 FPS while being ~100x faster to train over previous work.

Humans naturally interact with both others and the surrounding multiple objects, engaging in various social activities. However, recent advances in modeling human-object interactions mostly focus on perceiving isolated individuals and objects, due to fundamental data scarcity. In this paper, we introduce HOI-M$^3$, a novel large-scale dataset for modeling the interactions of Multiple huMans and Multiple objects. Notably, it provides accurate 3D tracking for both humans and objects from dense RGB and object-mounted IMU inputs, covering 199 sequences and 181M frames of diverse humans and objects under rich activities. With the unique HOI-M$^3$ dataset, we introduce two novel data-driven tasks with companion strong baselines: monocular capture and unstructured generation of multiple human-object interactions. Extensive experiments demonstrate that our dataset is challenging and worthy of further research about multiple human-object interactions and behavior analysis. Our HOI-M$^3$ dataset, corresponding codes, and pre-trained models will be disseminated to the community for future research.

We present InterHandGen, a novel framework that learns the generative prior of two-hand interaction. Sampling from our model yields plausible and diverse two-hand shapes in close interaction with or without an object. Our prior can be incorporated into any optimization or learning methods to reduce ambiguity in an ill-posed setup. Our key observation is that directly modeling the joint distribution of multiple instances imposes high learning complexity due to its combinatorial nature. Thus, we propose to decompose the modeling of joint distribution into the modeling of unconditional and conditional single instance distribution following Bayes' Theorem. In particular, we introduce a diffusion model that learns the single-hand distribution unconditional and conditional to another hand via conditioning dropout. For sampling, we combine anti-penetration and classifier-free guidance to enable plausible generation. Furthermore, we establish the rigorous evaluation protocol of two-hand synthesis, where our method significantly outperforms baseline generative models in terms of plausibility and diversity. We also demonstrate that our diffusion prior can boost the performance of two-hand reconstruction from monocular in-the-wild images, achieving new state-of-the-art accuracy. Our code will be publicly available.

A long-standing goal of 3D human reconstruction is to create lifelike and fully detailed 3D humans from single-view images. The main challenge lies in inferring unknown body shapes, appearances, and clothing details in areas not visible in the images. To address this, we propose SiTH, a novel pipeline that uniquely integrates an image-conditioned diffusion model into a 3D mesh reconstruction workflow. At the core of our method lies the decomposition of the challenging single-view reconstruction problem into generative hallucination and reconstruction subproblems. For the former, we employ a powerful generative diffusion model to hallucinate unseen back-view appearance based on the input images. For the latter, we leverage skinned body meshes as guidance to recover full-body texture meshes from the input and back-view images. SiTH requires as few as 500 3D human scans for training while maintaining its generality and robustness to diverse images. Extensive evaluations on two 3D human benchmarks, including our newly created one, highlighted our method's superior accuracy and perceptual quality in 3D textured human reconstruction.

The studies of human clothing for digital avatars have predominantly relied on synthetic datasets. While easy to collect, synthetic data often fall short in realism and fail to capture authentic clothing dynamics. Addressing this gap, we introduce 4D-DRESS, the first real-world 4D dataset advancing human clothing research with its high-quality 4D textured scans and garment meshes. 4D-DRESS captures 64 outfits in 520 human motion sequences, amounting to 78k textured scans. Creating a real-world clothing dataset is challenging, particularly in annotating and segmenting the extensive and complex 4D human scans. To address this, we develop a semi-automatic 4D human parsing pipeline. We efficiently combine a human-in-the-loop process with automation to accurately label 4D scans in diverse garments and body movements. Leveraging precise annotations and high-quality garment meshes, we establish several benchmarks for clothing simulation and reconstruction. 4D-DRESS offers realistic and challenging data that complements synthetic sources, paving the way for advancements in research of lifelike human clothing. Website: https://ait.ethz.ch/4d-dress.

We present MultiPly, a novel framework to reconstruct multiple people in 3D from monocular in-the-wild videos. Reconstructing multiple individuals moving and interacting naturally from monocular in-the-wild videos poses a challenging task. Addressing it necessitates precise pixel-level disentanglement of individuals without any prior knowledge about the subjects. Moreover, it requires recovering intricate and complete 3D human shapes from short video sequences, intensifying the level of difficulty. To tackle these challenges, we first define a layered neural representation for the entire scene, composited by individual human and background models. We learn the layered neural representation from videos via our layer-wise differentiable volume rendering. This learning process is further enhanced by our hybrid instance segmentation approach which combines the self-supervised 3D segmentation and the promptable 2D segmentation module, yielding reliable instance segmentation supervision even under close human interaction. A confidence-guided optimization formulation is introduced to optimize the human poses and shape/appearance alternately. We incorporate effective objectives to refine human poses via photometric information and impose physically plausible constraints on human dynamics, leading to temporally consistent 3D reconstructions with high fidelity. The evaluation of our method shows the superiority over prior art on publicly available datasets and in-the-wild videos.

The 3D Human Pose Estimation (3D HPE) task uses 2D images or videos to predict human joint coordinates in 3D space. Despite recent advancements in deep learning-based methods, they mostly ignore the capability of coupling accessible texts and naturally feasible knowledge of humans, missing out on valuable implicit supervision to guide the 3D HPE task. Moreover, previous efforts often study this task from the perspective of the whole human body, neglecting fine-grained guidance hidden in different body parts. To this end, we present a new Fine-Grained Prompt-Driven Denoiser based on a diffusion model for 3D HPE, named FinePOSE.It consists of three core blocks enhancing the reverse process of the diffusion model: (1) Fine-grained Part-aware Prompt learning (FPP) block constructs fine-grained part-aware prompts via coupling accessible texts and naturally feasible knowledge of body parts with learnable prompts to model implicit guidance. (2) Fine-grained Prompt-pose Communication (FPC) block establishes fine-grained communications between learned part-aware prompts and poses to improve the denoising quality. (3) Prompt-driven Timestamp Stylization (PTS) block integrates learned prompt embedding and temporal information related to the noise level to enable adaptive adjustment at each denoising step. Extensive experiments on public single-human pose estimation datasets show that FinePOSE outperforms state-of-the-art methods. We further extend FinePOSE to multi-human pose estimation. Achieving 34.3mm average MPJPE on the EgoHumans dataset demonstrates the potential of FinePOSE to deal with complex multi-human scenarios.Code is available at https://github.com/PKU-ICST-MIPL/FinePOSE_CVPR2024.

We present SimXR, a method for controlling a simulated avatar from information (headset pose and cameras) obtained from AR / VR headsets. Due to the challenging viewpoint of head-mounted cameras, the human body is often clipped out of view, making traditional image-based egocentric pose estimation challenging. On the other hand, headset poses provide valuable information about overall body motion, but lack fine-grained details about the hands and feet. To synergize headset poses with cameras, we control a humanoid to track headset movement while analyzing input images to decide body movement. When body parts are seen, the movements of hands and feet will be guided by the images; when unseen, the laws of physics guide the controller to generate plausible motion. We design an end-to-end method that does not rely on any intermediate representations and learns to directly map from images and headset poses to humanoid control signals. To train our method, we also propose a large-scale synthetic dataset created using camera configurations compatible with a commercially available VR headset (Quest 2) and show promising results on real-world captures. To demonstrate the applicability of our framework, we also test it on an AR headset with a forward-facing camera.

In this work, we present Digital Life Project, a framework utilizing language as the universal medium to build autonomous 3D characters, who are capable of engaging in social interactions and expressing with articulated body motions, thereby simulating life in a digital environment. Our framework comprises two primary components: 1) SocioMind: a meticulously crafted digital brain that models personalities with systematic few-shot exemplars, incorporates a reflection process based on psychology principles, and emulates autonomy by initiating dialogue topics; 2) MoMat-MoGen: a text-driven motion synthesis paradigm for controlling the character’s digital body. It integrates motion matching, a proven industry technique to ensure motion quality, with cutting-edge advancements in motion generation for diversity. Extensive experiments demonstrate that each module achieves state-of-the-art performance in its respective domain. Collectively, they enable virtual characters to initiate and sustain dialogues autonomously, while evolving their socio-psychological states. Concurrently, these characters can perform contextually relevant bodily movements. Additionally, an extension of DLP enables a virtual character to recognize and appropriately respond to human players’ actions.

Gait, a prevalent and complex form of human motion, plays a significant role in the field of long-range pedestrian retrieval due to the unique characteristics inherent in individual motion patterns. However, gait recognition in real-world scenarios is challenging due to the limitations of capturing comprehensive cross-viewing and cross-clothing data. Additionally, distractors such as occlusions, directional changes, and lingering movements further complicate the problem. The widespread application of deep learning techniques has led to the development of various potential gait recognition methods. However, these methods utilize convolutional networks to extract shared information across different views and attire conditions. Once trained, the parameters and non-linear function become constrained to fixed patterns, limiting their adaptability to various distractors in real-world scenarios. In this paper, we present a unified gait recognition framework to extract global motion patterns and develop a novel dynamic transformer to generate representative gait features. Specifically, we develop a trainable part-based prompt pool with numerous key-value pairs that can dynamically select prompt templates to incorporate into the gait sequence, thereby providing task-relevant shared knowledge information. Furthermore, we specifically design dynamic attention to extract robust motion patterns and address the length generalization issue. Extensive experimentation on four widely recognized gait datasets, i.e., Gait3D, GREW, OUMVLP, and CASIA-B, reveals that the proposed method yields substantial improvements compared to current state-of-the-art approaches.

Transformers have been successfully applied in the field of video-based 3D human pose estimation. However, the high computational costs of these video pose transformers (VPTs) make them impractical on resource-constrained devices. In this paper, we present a plug-and-play pruning-and-recovering framework, called Hourglass Tokenizer (HoT), for efficient transformer-based 3D human pose estimation from videos. Our HoT begins with pruning pose tokens of redundant frames and ends with recovering full-length tokens, resulting in a few pose tokens in the intermediate transformer blocks and thus improving the model efficiency. To effectively achieve this, we propose a token pruning cluster (TPC) that dynamically selects a few representative tokens with high semantic diversity while eliminating the redundancy of video frames. In addition, we develop a token recovering attention (TRA) to restore the detailed spatio-temporal information based on the selected tokens, thereby expanding the network output to the original full-length temporal resolution for fast inference. Extensive experiments on two benchmark datasets (i.e., Human3.6M and MPI-INF-3DHP) demonstrate that our method can achieve both high efficiency and estimation accuracy compared to the original VPT models. For instance, applying to MotionBERT and MixSTE on Human3.6M, our HoT can save nearly 50\% FLOPs without sacrificing accuracy and nearly 40% FLOPs with only 0.2% accuracy drop, respectively. Code and models are available at https://github.com/NationalGAILab/HoT.

The capacity of existing human keypoint localization models is limited by keypoint priors provided by the training data. To alleviate this restriction and pursue more general model, this work studies keypoint localization from a different perspective by reasoning locations based on keypiont clues in text descriptions. We propose LocLLM, the first Large-Language Model (LLM) based keypoint localization model that takes images and text instructions as inputs and outputs the desired keypoint coordinates.LocLLM leverages the strong reasoning capability of LLM and clues of keypoint type, location, and relationship in textual descriptions for keypoint localization. To effectively tune LocLLM, we construct localization-based instruction conversations to connect keypoint description with corresponding coordinates in input image, and fine-tune the whole model in a parameter-efficient training pipeline. LocLLM shows remarkable performance on standard 2D/3D keypoint localization benchmarks. Moreover, incorporating language clues into the localization makes LocLLM show superior flexibility and generalizable capability in cross dataset keypoint localization, and even detecting novel type of keypoints unseen during training. We will release the model and code for further research and evaluation.

Regression-based keypoint localization shows advantages of high efficiency and better robustness to quantization errors than heatmap-based methods. However, existing regression-based methods discard the spatial location prior in input image with a global pooling, leading to inferior accuracy and are limited to single instance localization tasks. We study the regression-based keypoint localization from a new perspective by leveraging the spatial location prior. Instead of regressing on the pooled feature, the proposed Spatial-Aware Regression (SAR) maintains the spatial location map and outputs spatial coordinates and confidence score for each grid, which are optimized with a unified objective. Benefited by the location prior, these spatial-aware outputs can be efficiently optimized, resulting in better localization performance. Moreover, incorporating spatial prior makes SAR more general and can be applied into various keypoint localization tasks. We test the proposed method in 4 keypoint localization tasks including single/multi-person 2D/3D pose estimation, and the whole-body pose estimation. Extensive experiments demonstrate its promising performance, e.g., consistently outperforming recent regressions-based methods.

We present GaussianAvatar, an efficient approach to creating realistic human avatars with dynamic 3D appearances from a single video.We start by introducing animatable 3D Gaussians to explicitly represent humans in various poses and clothing styles.Such an explicit and animatable representation can fuse 3D appearances more efficiently and consistently from 2D observations.Our representation is further augmented with dynamic properties to support pose-dependent appearance modeling, where a dynamic appearance network along with an optimizable feature tensor is designed to learn the motion-to-appearance mapping.Moreover, by leveraging the differentiable motion condition, our method enables a joint optimization of motions and appearances during avatar modeling, which helps to tackle the long-standing issue of inaccurate motion estimation in monocular settings.The efficacy of GaussianAvatar is validated on both the public dataset and our collected dataset, demonstrating its superior performances in terms of appearance quality and rendering efficiency.

Recent years have witnessed a trend of the deep integration of the generation and reconstruction paradigms. In this paper, we extend the ability of controllable generative models for a more comprehensive hand mesh recovery task: direct hand mesh generation, inpainting, reconstruction, and fitting in a single framework, which we name as Holistic Hand Mesh Recovery (HHMR).Our key observation is that different kinds of hand mesh recovery tasks can be achieved by a single generative model with strong multimodal controllability, and in such a framework, realizing different tasks only requires giving different signals as conditions. To achieve this goal, we propose an all-in-one diffusion framework based on graph convolution and attention mechanisms for holistic hand mesh recovery. In order to achieve strong control generation capability while ensuring the decoupling of multimodal control signals, we map different modalities to a share feature space and apply cross-scale random masking in both modality and feature levels. In this way, the correlation between different modalities can be fully exploited during the learning of hand priors. Furthermore, we propose Condition-aligned Gradient Guidance to enhance the alignment of the generated model with the control signals, which significantly improves the accuracy of the hand mesh reconstruction and fitting. Experiments show that our novel framework can realize multiple hand mesh recovery tasks simultaneously and outperform the existing methods in different tasks, which provides more possibilities for subsequent downstream applications including gesture recognition, pose generation, mesh editing, and so on.

In this paper, we focus on capturing closely interacted two-person motions from monocular videos, an important yet understudied topic. Unlike less-interacted motions, closely interacted motions contain frequently occurring inter-human occlusions, which pose significant challenges to existing capturing algorithms. To address this problem, our key observation is that close physical interactions between two subjects typically happen under very specific situations (e.g., handshake, hug, etc.), and such situational contexts contain strong prior semantics to help infer the poses of occluded joints. In this spirit, we introduce reaction priors, which are invertible neural networks that bi-directionally model the pose probability distributions of one person given the pose of the other. The learned reaction priors are then incorporated into a query-based pose estimator, which is a decoder-only Transformer with self-attentions on both intra-joint and inter-joint relationships. We demonstrate that our design achieves considerably higher performance than previous methods on multiple benchmarks. What's more, as existing datasets lack sufficient cases of close human-human interactions, we also build a new dataset called Dual-Human to better evaluate different methods. Dual-Human contains around $2k$ sequences of closely interacted two-person motions, each with synthetic multi-view renderings, contact annotations, and text descriptions. We believe that this new public dataset can significantly promote further research in this area.

Achieving high synchronization in the synthesis of realistic, speech-driven talking head videos presents a significant challenge. Traditional Generative Adversarial Networks (GAN) struggle to maintain consistent facial identity, while Neural Radiance Fields (NeRF) methods, although they can address this issue, often produce mismatched lip movements, inadequate facial expressions, and unstable head poses. A lifelike talking head requires synchronized coordination of subject identity, lip movements, facial expressions, and head poses. The absence of these synchronizations is a fundamental flaw, leading to unrealistic and artificial outcomes. To address the critical issue of synchronization, identified as the “devil” in creating realistic talking heads, we introduce SyncTalk. This NeRF-based method effectively maintains subject identity, enhancing synchronization and realism in talking head synthesis. SyncTalk employs a Face-Sync Controller to align lip movements with speech and innovatively uses a 3D facial blendshape model to capture accurate facial expressions. Our Head-Sync Stabilizer optimizes head poses, achieving more natural head movements. The Portrait-Sync Generator restores hair details and blends the generated head with the torso for a seamless visual experience. Extensive experiments and user studies demonstrate that SyncTalk outperforms state-of-the-art methods in synchronization and realism. Our code will be publicly available.

The pursuit of accurate 3D hand pose estimation stands as a keystone for understanding human activity in the realm of egocentric vision. The majority of existing estimation methods still rely on single-view images as input, leading to potential limitations, e.g., limited field-of-view and ambiguity in depth. To address these problems, adding another camera to better capture the shape of hands is a practical direction. However, existing multi-view hand pose estimation methods suffer from two main drawbacks: 1) Requiring multi-view annotations for training, which are expensive. 2) During testing, the model becomes inapplicable if camera parameters/layout are not the same as those used in training. In this paper, we propose a novel Single-to-Dual-view adaptation (S2DHand) solution that adapts a pre-trained single-view estimator to dual views. Compared with existing multi-view training methods, 1) our adaptation process is unsupervised, eliminating the need for multi-view annotation. 2) Moreover, our method can handle arbitrary dual-view pairs with unknown camera parameters, making the model applicable to diverse camera settings. Specifically, S2DHand is built on certain stereo constraints, including pair-wise cross-view consensus and invariance of transformation between both views. These two stereo constraints are used in a complementary manner to generate pseudo-labels, allowing reliable adaptation. Evaluation results reveal that S2DHand achieves significant improvements on arbitrary camera pairs under both in-dataset and cross-dataset settings, and outperforms existing adaptation methods with leading performance.

Dance serves as a powerful medium for expressing human emotions, but the lifelike generation of dance is still a considerable challenge.Recently, diffusion models have showcased remarkable generative abilities across various domains. They hold promise for human motion generation due to their adaptable many-to-many nature. Nonetheless, current diffusion-based motion generation models often create entire motion sequences directly and unidirectionally, lacking focus on the motion with local and bidirectional enhancement. When choreographing high-quality dance movements, people need to take into account not only the musical context but also the nearby music-aligned dance motions. To authentically capture human behavior, we propose a Bidirectional Autoregressive Diffusion Model (BADM) for music-to-dance generation, where a bidirectional encoder is built to enforce that the generated dance is harmonious in both the forward and backward directions. To make the generated dance motion smoother, a local information decoder is built for local motion enhancement. The proposed framework is able to generate new motions based on the input conditions and nearby motions, which foresees individual motion slices iteratively and consolidates all predictions. To further refine the synchronicity between the generated dance and the beat, the beat information is incorporated as an input to generate better music-aligned dance movements. Experimental results demonstrate that the proposed model achieves state-of-the-art performance compared to existing unidirectional approaches on the prominent benchmark for music-to-dance generation.

Facial geometry and appearance capture have demonstrated tremendous success in 3D scanning real humans in studios. Recent works propose to democratize this technique while keeping the results high quality. However, they are still inconvenient for daily usage. In addition, they focus on an easier problem of only capturing facial skin. This paper proposes a novel method for high-quality face capture, featuring an easy-to-use system and the capability to model the complete face with skin, mouth interior, hair, and eyes. We reconstruct facial geometry and appearance from a single co-located smartphone flashlight sequence captured in a dim room where the flashlight is the dominant light source (\emph{e.g.} rooms with curtains or at night). To model the complete face, we propose a novel hybrid representation to effectively model both eyes and other facial regions, along with novel techniques to learn it from images. We apply a combined lighting model to compactly represent real illuminations and exploit a morphable face albedo model as a reflectance prior to disentangle diffuse and specular. Experiments show that our method can capture high-quality 3D relightable scans. Our code will be released.

In this work, we aim to improve the 3D reasoning ability of Transformers in multi-view 3D human pose estimation. Recent works have focused on end-to-end learning-based transformer designs, which struggle to resolve geometric information accurately, particularly during occlusion. Instead, we propose a novel hybrid model, MVGFormer, which has a series of geometric and appearance modules organized in an iterative manner. The geometry modules are learning-free and handle all viewpoint-dependent 3D tasks geometrically which notably improves the model's generalization ability. The appearance modules are learnable and are dedicated to estimating 2D poses from image signals end-to-end which enables them to achieve accurate estimates even when occlusion occurs, leading to a model that is both accurate and generalizable to new cameras and geometries. We evaluate our approach for both in-domain and out-of-domain settings, where our model consistently outperforms state-of-the-art methods, and especially does so by a significant margin in the out-of-domain setting. We will release the code and models: https://github.com/XunshanMan/MVGFormer.

We address the challenge of content diversity and controllability in pedestrian simulation for driving scenarios. Recent pedestrian animation frameworks have a significant limitation wherein they primarily focus on either the following trajectory or the content of the reference video, consequently overlooking the potential diversity of human motion within such scenarios. This limitation restricts the ability to generate pedestrian behaviors that exhibit a wider range of variations and realistic motions and therefore restricts its usage to provide rich motion content for other components in the driving simulation system, eg., suddenly changed motion to which the autonomous vehicle should respond. In our approach, we strive to surpass the limitation by showcasing diverse human motions obtained from various sources, such as generated human motions, in addition to following the given trajectory. The fundamental contribution of our framework lies in combining the motion tracking task with trajectory following, which enables the tracking of specific motion parts (eg., upper body) while simultaneously following the given trajectory by a single policy. This way, we significantly enhance both the diversity of simulated human motion within the given scenario and the controllability of the content, including language-based control. Our framework facilitates the generation of a wide range of human motions, contributing to greater realism and adaptability in pedestrian simulations for driving scenarios.

We are living in a world surrounded by diverse and “smart” devices with rich modalities of sensing ability. Conveniently capturing the interactions between us humans and these objects remains far-reaching. In this paper, we present I'm-HOI, a monocular scheme to faithfully capture the 3D motions of both the human and object in a novel setting: using a minimal amount of RGB camera and object-mounted Inertial Measurement Unit (IMU). It combines general motion inference and category-aware refinement. For the former, we introduce a holistic human-object tracking method to fuse the IMU signals and the RGB stream and progressively recover the human motions and subsequently the companion object motions. For the latter, we tailor a category-aware motion diffusion model, which is conditioned on both the raw IMU observations and the results from the previous stage under over-parameterization representation. It significantly refines the initial results and generates vivid body, hand, and object motions. Moreover, we contribute a large dataset with ground truth human and object motions, dense RGB inputs, and rich object-mounted IMU measurements. Extensive experiments demonstrate the effectiveness of I'm-HOI under a hybrid capture setting. Our dataset and code will be released to the community.

For human avatar reconstruction, contemporary techniques commonly necessitate the acquisition of costly data and struggle to achieve satisfactory results from a small number of casual images. In this paper, we investigate this task from a few-shot unconstrained photo album. The reconstruction of human avatars from such data sources is challenging because of limited data amount and dynamic articulated poses. For handling dynamic data, we integrate a skinning mechanism with deep marching tetrahedra (DMTet) to form a drivable tetrahedra representation, which drives arbitrary mesh topologies generated by the DMTet for the adaptation of unconstrained images. To effectively mine instructive information from few-shot data, we devise a two-phase optimization method with few-shot reference and few-shot guidance. The former focuses on aligning avatar identity with reference images, while the latter aims to generate plausible appearances for unseen regions. Overall, our framework, called HaveFun, can undertake avatar reconstruction, rendering, and animation. Extensive experiments on our developed benchmarks demonstrate that HaveFun exhibits substantially superior performance in reconstructing the human body and hand.

Language models have demonstrated impressive ability in context understanding and generative performance. Inspired by the recent success of language foundation models, in this paper, we propose LMTraj (Language-based Multimodal Trajectory predictor), which recasts the trajectory prediction task into a sort of question-answering problem. Departing from traditional numerical regression models, which treat the trajectory coordinate sequence as continuous signals, we consider them as discrete signals like text prompts. Specially, we first transform an input space for the trajectory coordinate into the natural language space. Here, the entire time-series trajectories of pedestrians are converted into a text prompt, and scene images are described as text information through image captioning. The transformed numerical and image data are then wrapped into the question-answering template for use in a language model. Next, to guide the language model in understanding and reasoning high-level knowledge, such as scene context and social relationships between pedestrians, we introduce an auxiliary multi-task question and answering. We then train a numerical tokenizer with the prompt data. We encourage the tokenizer to separate the integer and decimal parts well, and leverage it to capture correlations between the consecutive numbers in the language model. Lastly, we train the language model using the numerical tokenizer and all of the question-answer prompts. Here, we propose a beam-search-based most-likely prediction and a temperature-based multimodal prediction to implement both deterministic and stochastic inferences. Applying our LMTraj, we show that the language-based model can be a powerful pedestrian trajectory predictor, and outperforms existing numerical-based predictor methods. Extensive experiments show that our LMTraj can successfully understand social relationships and accurately extrapolate the multimodal futures on the public pedestrian trajectory prediction benchmark. Code is publicly available at https://github.com/InhwanBae/LMTrajectory.

While head-mounted devices are becoming more compact, they provide egocentric views with significant self-occlusions of the device user. Hence, existing methods often fail to accurately estimate complex 3D poses from egocentric views. In this work, we propose a new transformer-based framework to improve egocentric stereo 3D human pose estimation, which leverages the scene information and temporal context of egocentric stereo videos. Specifically, we utilize 1) depth features from our 3D scene reconstruction module with uniformly sampled windows of egocentric stereo frames, and 2) human joint queries enhanced by temporal features of the video inputs. Our method is able to accurately estimate human poses even in challenging scenarios, such as crouching and sitting. Furthermore, we introduce two new benchmark datasets, i.e., UnrealEgo2 and UnrealEgo-RW (Real World). The proposed datasets offer a much larger number of egocentric stereo views with a wider variety of human motions than the existing datasets, allowing comprehensive evaluation of existing and upcoming methods. Our extensive experiments show that the proposed approach significantly outperforms previous methods. We will release UnrealEgo2, UnrealEgo-RW, and trained models on our project page.

In this work, we explore egocentric whole-body motion capture using a single fisheye camera, which simultaneously estimates human body and hand motion. This task presents significant challenges due to three factors: the lack of high-quality datasets, fisheye camera distortion, and human body self-occlusion. To address these challenges, we propose a novel approach that leverages FisheyeViT to extract fisheye image features, which are subsequently converted into pixel-aligned 3D heatmap representations for 3D human body pose prediction. For hand tracking, we incorporate dedicated hand detection and hand pose estimation networks for regressing 3D hand poses. Finally, we develop a diffusion-based whole-body motion prior model to refine the estimated whole-body motion while accounting for joint uncertainties. To train these networks, we collect a large synthetic dataset, EgoWholeBody, comprising 840,000 high-quality egocentric images captured across a diverse range of whole-body motion sequences. Quantitative and qualitative evaluations demonstrate the effectiveness of our method in producing high-quality whole-body motion estimates from a single egocentric camera.

This work addresses the problem of real-time rendering of photorealistic human body avatars learned from multi-view videos. While the classical approaches to model and render virtual humans generally use a textured mesh, recent research has developed neural body representations that achieve impressive visual quality. However, these models are difficult to render in real-time and their quality degrades when the character is animated with body poses different than the training observations. We propose an animatable human model based on 3D Gaussian Splatting, that has recently emerged as a very efficient alternative to neural radiance fields. The body is represented by a set of gaussian primitives in a canonical space which is deformed with a coarse to fine approach that combines forward skinning and local non-rigid refinement. We describe how to learn our Human Gaussian Splatting (HuGS) model in an end-to-end fashion from multi-view observations, and evaluate it against the state-of-the-art approaches for novel pose synthesis of clothed body. Our method achieves 1.5 dB PSNR improvement over the state-of-the-art on THuman4 dataset while being able to render in real-time (≈ 80 fps for 512 × 512 resolution).

In this paper, we delve into the creation of one-shot hand avatars, attaining high-fidelity and drivable hand representations swiftly from a single image. With the burgeoning domains of the digital human, the need for quick and personalized hand avatar creation has become increasingly critical. Existing techniques typically require extensive input data and may prove cumbersome or even impractical in certain scenarios. To enhance accessibility, we present a novel method OHTA (One-shot Hand avaTAr) that enables the creation of detailed hand avatars from merely one image. OHTA tackles the inherent difficulties of this data-limited problem by learning and utilizing data-driven hand priors. Specifically, we design a hand prior model initially employed for 1) learning various hand priors with available data and subsequently for 2) the inversion and fitting of the target identity with prior knowledge. OHTA demonstrates the capability to create high-fidelity hand avatars with consistent animatable quality, solely relying on a single image. Furthermore, we illustrate the versatility of OHTA through diverse applications, encompassing text-to-avatar conversion, hand editing, and identity latent space manipulation.

To date, the quest to rapidly and effectively produce human-object interaction (HOI) animations directly from textual descriptions stands at the forefront of computer vision research. The underlying challenge demands both a discriminating interpretation of language and a comprehensive physics-centric model supporting real-world dynamics. To ameliorate, this paper advocates HOIAnimator, a novel and interactive diffusion model with perception ability and also ingeniously crafted to revolutionize the animation of complex interactions from linguistic narratives. The effectiveness of our model is anchored in two ground-breaking innovations: (1) Our Perceptive Diffusion Models (PDM) brings together two types of models: one focused on human movements and the other on objects. This combination allows for animations where humans and objects move in concert with each other, making the overall motion more realistic. Additionally, we propose a Perceptive Message Passing (PMP) mechanism to enhance the communication bridging the two models, ensuring that the animations are smooth and unified; (2) We devise an Interaction Contact Field (ICF), a sophisticated model that implicitly captures the essence of HOIs. Beyond mere predictive contact points, the ICF assesses the proximity of human and object to their respective environment, informed by a probabilistic distribution of interactions learned throughout the denoising phase. Our comprehensive evaluation showcases HOIanimator's superior ability to produce dynamic, context-aware animations that surpass existing benchmarks in text-driven animation synthesis.

Computer animation's quest to bridge content and style has historically been a challenging venture, with previous efforts often leaning toward one at the expense of the other. This paper tackles the inherent challenge of content-style duality, ensuring a harmonious fusion where the core narrative of the content is both preserved and elevated through stylistic enhancements. We propose a novel Multi-condition Motion Latent Diffusion Model (MCM-LDM) for Arbitrary Motion Style Transfer (AMST). Our MCM-LDM significantly emphasizes preserving trajectories, recognizing their fundamental role in defining the essence and fluidity of motion content. Our MCM-LDM's cornerstone lies in its ability first to disentangle and then intricately weave together motion's tripartite components: motion trajectory, motion content, and motion style. The critical insight of MCM-LDM is to embed multiple conditions with distinct priorities. The content channel serves as the primary flow, guiding the overall structure and movement, while the trajectory and style channels act as auxiliary components and synchronize with the primary one dynamically. This mechanism ensures that multi-conditions can seamlessly integrate into the main flow, enhancing the overall animation without overshadowing the core content. Empirical evaluations underscore the model's proficiency in achieving fluid and authentic motion style transfers, setting a new benchmark in the realm of computer animation. Upon its final acceptance, our source codes will be made publicly available.

In this work, we explore a novel task of generating human grasps based on single-view scene point clouds, which more accurately mirrors the typical real-world situation of observing objects from a single viewpoint. Due to the incompleteness of object point clouds and the presence of numerous scene points, the generated hand is prone to penetrating into the invisible parts of the object and the model is easily affected by scene points. Thus, we introduce S2HGrasp, a framework composed of two key modules: the Global Perception module that globally perceives partial object point clouds, and the DiffuGrasp module designed to generate high-quality human grasps based on complex inputs that include scene points. Additionally, we introduce S2HGD dataset, which comprises approximately 99,000 single-object single-view scene point clouds of 1,668 unique objects, each annotated with one human grasp. Our extensive experiments demonstrate that S2HGrasp can not only generate natural human grasps regardless of scene points, but also effectively prevent penetration between the hand and invisible parts of the object. Moreover, our model showcases strong generalization capability when applied to unseen objects. Our code and dataset are available at https://github.com/iSEE-Laboratory/S2HGrasp.

We present EgoTAP, a heatmap-to-3D pose lifting method for highly accurate stereo egocentric 3D pose estimation. Severe self-occlusion and out-of-view limbs in egocentric camera views make accurate pose estimation a challenging problem. To address the challenge, prior methods employ joint heatmaps-probabilistic 2D representations of the body pose, but heatmap-to-3D pose conversion still remains an inaccurate process. We propose a novel heatmap-to-3D lifting method composed of the Grid ViT Encoder and the Propagation Network. The Grid ViT Encoder summarizes joint heatmaps into effective feature embedding using self-attention. Then, the Propagation Network estimates the 3D pose by utilizing skeletal information to better estimate the position of obscure joints. Our method significantly outperforms the previous state-of-the-art qualitatively and quantitatively demonstrated by a 23.9\% reduction of error in an MPJPE metric. Our source code is available on GitHub.

Existing photorealistic relightable hand models require extensive identity-specific observations in different views, poses, and illuminations, and face challenges in generalizing to natural illuminations and novel identities. To bridge this gap, we present URHand, the first universal relightable hand model that generalizes across viewpoints, poses, illuminations, and identities. Our model allows few-shot personalization using images captured with a mobile phone, and is ready to be photorealistically rendered under novel illuminations. To simplify the personalization process while retaining photorealism, we build a powerful universal relightable prior based on neural relighting from multi-view images of hands captured in a light stage with hundreds of identities. The key challenge is scaling the cross-identity training while maintaining personalized fidelity and sharp details without compromising generalization under natural illuminations. To this end, we propose a spatially varying linear lighting model as the neural renderer that takes physics-inspired shading as input feature. By removing non-linear activations and bias, our specifically designed lighting model explicitly keeps the linearity of light transport. This enables single-stage training from light-stage data while generalizing to real-time rendering under arbitrary continuous illuminations across diverse identities. In addition, we introduce the joint learning of a physically based model and our neural relighting model, which further improves fidelity and generalization. Extensive experiments show that our approach achieves superior performance over existing methods in terms of both quality and generalizability. We also demonstrate quick personalization of our universal relightable model from a short phone scan of an unseen identity.

Existing physics-based motion generation methods rely heavily on imitation learning and reward shaping to excel in motion quality and interactivity, which hinders their ability to generalize to unseen scenarios. To address this limitation, we propose AnySkill, a novel hierarchical method that can learn physically plausible interactions following open-vocabulary instructions. Our approach first constructs a repertoire of atomic actions by learning a low-level controller through imitation learning. Then, given an open-vocabulary text instruction, we train a high-level policy that assembles appropriate atomic actions to maximize the CLIP similarity between the rendered images of the agent and the text instruction. Moreover, since we employ an image-based reward design for our high-level policy, the agent can naturally learn the interactions with objects without hand-crafted reward engineering. We show the ability of AnySkill to learn reasonable and natural motion sequences in response to unseen captions of varying lengths, demonstrating the first method to perform open-vocabulary physical skill for interactive humanoid agents.

We tackle a new problem of multi-view camera and subject registration in the bird's eye view (BEV) without pre-given camera calibration, which promotes the multi-view subject registration problem to a new calibration-free stage. This greatly alleviates the limitation in many practical applications. However, this is a very challenging problem since its only input is several RGB images from different first-person views (FPVs) for a multi-person scene, without the BEV image and the calibration of the FPVs, while the output is a unified plane aggregated from all views with the positions and orientations of both the subjects and cameras in a BEV. For this purpose, we propose an end-to-end framework solving camera and subject registration together by taking advantage of their mutual dependence, whose main idea is as below: i) creating a subject view-transform module (VTM) to project each pedestrian from FPV to a virtual BEV, ii) deriving a multi-view geometry-based spatial alignment module (SAM) to estimate the relative camera pose in a unified BEV, iii) selecting and refining the subject and camera registration results within the unified BEV. We collect a new large-scale synthetic dataset with rich annotations for training and evaluation. Additionally, we also collect a real dataset for cross-domain evaluation. The experimental results show the remarkable effectiveness of our method. We will release the dataset and code to the public.

It is especially challenging to achieve real-time human motion tracking on a standalone VR Head-Mounted Display (HMD) such as Meta Quest and PICO. In this paper, we propose HMD-Poser, the first unified approach to recover full-body motions using scalable sparse observations from HMD and body-worn IMUs. In particular, it can support a variety of input scenarios, such as HMD, HMD+2IMUs, HMD+3IMUs, etc. The scalability of inputs may accommodate users' choices for both high tracking accuracy and easy-to-wear. A lightweight temporal-spatial feature learning network is proposed in HMD-Poser to guarantee that the model runs in real-time on HMDs. Furthermore, HMD-Poser presents online body shape estimation to improve the position accuracy of body joints. Extensive experimental results on the challenging AMASS dataset show that HMD-Poser achieves new state-of-the-art results in both accuracy and real-time performance. We also build a new free-dancing motion dataset to evaluate HMD-Poser's on-device performance and investigate the performance gap between synthetic data and real-captured sensor data. Finally, we demonstrate our HMD-Poser with a real-time Avatar-driving application on a commercial HMD. Our code and free-dancing motion dataset are available \href{https://pico-ai-team.github.io/hmd-poser}{here}.

Recent 3D face reconstruction methods have made remarkable advancements, yet there remain huge challenges in monocular high-quality facial reflectance reconstruction. Existing methods rely on a large amount of light-stage captured data to learn facial reflectance models. However, the lack of subject diversity poses challenges in achieving good generalization and widespread applicability. In this paper, we learn the reflectance prior in image space rather than UV space and present a framework named ID2Reflectance. Our framework can directly estimate the reflectance maps of a single image while using limited captured data for training. Our key insight is that reflectance data shares facial structures with RGB faces, which enables obtaining expressive facial prior from inexpensive RGB data and reduces the dependency on reflectance data. We first learn a high-quality prior for facial reflectance. Specifically, we pretrain multi-domain facial feature codebooks and design a codebook fusion method to align the reflectance and RGB domains. Then, we propose an identity-conditioned swapping module that injects facial identity from the target image into the pre-trained autoencoder to modify the identity of the original reflectance image. Finally, we stitch swapped reflectance images in wrapped space to obtain renderable assets. Extensive experiments demonstrate that our method exhibits excellent generalization capability and achieves state-of-the-art facial reflectance reconstruction results for in-the-wild faces.

Gaussian splatting has emerged as a powerful 3D representation that harnesses the advantages of both explicit (mesh) and implicit (NeRF) 3D representations. In this paper, we seek to leverage Gaussian splatting to generate realistic animatable avatars from textual descriptions, addressing the limitations (e.g., efficiency and flexibility) imposed by mesh or NeRF-based representations. However, a naive application of Gaussian splatting cannot generate high-quality animatable avatars and suffers from learning instability; it also cannot capture fine avatar geometries and often leads to degenerate body parts. To tackle these problems, we first propose a primitive-based 3D Gaussian representation where Gaussians are defined inside pose-driven primitives to facilitate animations. Second, to stabilize and amortize the learning of millions of Gaussians, we propose to use implicit neural fields to predict the Gaussian attributes (e.g., colors). Finally, to capture fine avatar geometries and extract detailed meshes, we propose a novel SDF-based implicit mesh learning approach for 3D Gaussians that regularizes the underlying geometries and extracts highly detailed textured meshes. Our proposed method, GAvatar, enables the large-scale generation of diverse animatable avatars using only text prompts. GAvatar significantly surpasses existing methods in terms of both appearance and geometry quality, and achieves extremely fast rendering (100 fps) at 1K resolution.

We present Score-Guided Human Mesh Recovery (ScoreHMR), an approach for solving inverse problems for 3D human pose and shape reconstruction. These inverse problems involve fitting a human body model to image observations, traditionally solved through optimization techniques. ScoreHMR mimics model fitting approaches, but alignment with the image observation is achieved through score guidance in the latent space of a diffusion model. The diffusion model is trained to capture the conditional distribution of the human model parameters given an input image. By guiding its denoising process with a task-specific score, ScoreHMR effectively solves inverse problems for various applications without the need for retraining the task-agnostic diffusion model. We evaluate our approach on three settings/applications. These are: (i) single-frame model fitting; (ii) reconstruction from multiple uncalibrated views; (iii) reconstructing humans in video sequences. ScoreHMR consistently outperforms all optimization baselines on popular benchmarks across all settings. We make our code and models available on the project website: https://statho.github.io/ScoreHMR.

3D facial editing, a longstanding task in computer vision with broad applications, is expected to fast and intuitively manipulate any face from arbitrary viewpoints following the user's will. Existing works have limitations in terms of intuitiveness, generalization, and efficiency. To overcome these challenges, we propose FaceEdit3D, which allows users to directly manipulate 3D points to edit a 3D face, achieving natural and rapid face editing. After one or several points are manipulated by users, we propose the tri-plane warping to directly manipulate the view-independent 3D representation. To address the problem of distortion caused by tri-plane warping, we train a warp-aware encoder to project the warped face onto a standardized latent space. In this space, we further propose directional latent editing to mitigate the identity bias caused by the encoder and realize the disentangled editing of various attributes. Extensive experiments show that our method achieves superior results with rich facial details and nice identity preservation. Our approach also supports general applications like multi-attribute continuous editing and car editing.

Synthesizing natural human motions that enable a 3D human avatar to walk and reach for arbitrary goals in 3D space remains an unsolved problem with many applications. Existing methods (data-driven or using reinforcement learning) are limited in terms of generalization and motion naturalness. A primary obstacle is the scarcity of training data that combines locomotion with goal reaching. To address this, we introduce WANDR, a data-driven model that takes an avatar’s initial pose and a goal’s 3D position and generates natural human motions that place the end effector (wrist) on the goal location. To solve this, we introduce novel intention features that drive rich goal-oriented movement. Intention guides the agent to the goal, and interactively adapts the generation to novel situations without needing to define sub-goals or the entire motion path. Crucially, intention allows training on datasets that have goal-oriented motions as well as those that do not. WANDR is a conditional Variational Auto-Encoder (c-VAE), which we train using the AMASS and CIRCLE datasets. We evaluate our method extensively and demonstrate its ability to generate natural and long-term motions that reach 3D goals and generalize to unseen goal locations. Our models and code are available for research purposes at wandr.is.tue.mpg.de

To build a cross-modal latent space between 3D human motion and language, acquiring large-scale and high-quality human motion data is crucial. However, unlike the abundance of image data, the scarcity of motion data has limited the performance of existing motion-language models. To counter this, we introduce ``motion patches'', a new representation of motion sequences, and propose using Vision Transformers (ViT) as motion encoders via transfer learning, aiming to extract useful knowledge from the image domain and apply it to the motion domain. These motion patches, created by dividing and sorting skeleton joints based on body parts in motion sequences, are robust to varying skeleton structures, and can be regarded as color image patches in ViT. We find that transfer learning with pre-trained weights of ViT obtained through training with 2D image data can boost the performance of motion analysis, presenting a promising direction for addressing the issue of limited motion data. Our extensive experiments show that the proposed motion patches, used jointly with ViT, achieve state-of-the-art performance in the benchmarks of text-to-motion retrieval, and other novel challenging tasks, such as cross-skeleton recognition, zero-shot motion classification, and human interaction recognition, which are currently impeded by the lack of data.

We address the problem of generating realistic 3D motions of humans interacting with objects in a scene. Our key idea is to create a neural interaction field attached to a specific object, which outputs the distance to the valid interaction manifold given a human pose as input. This interaction field guides the sampling of an object-conditioned human motion diffusion model, so as to encourage plausible contacts and affordance semantics. To support interactions with scarcely available data, we propose an automated synthetic data pipeline. For this, we seed a pre-trained motion model, which has priors for the basics of human movement, with interaction-specific anchor poses extracted from limited motion capture data. Using our guided diffusion model trained on generated synthetic data, we synthesize realistic motions for sitting and lifting with several objects, outperforming alternative approaches in terms of motion quality and successful action completion. We call our framework NIFTY: Neural Interaction Fields for Trajectory sYnthesis

We present DreamAvatar, a text-and-shape guided framework for generating high-quality 3D human avatars with controllable poses. While encouraging results have been reported by recent methods on text-guided 3D common object generation, generating high-quality human avatars remains an open challenge due to the complexity of the human body's shape, pose, and appearance. We propose DreamAvatar to tackle this challenge, which utilizes a trainable NeRF for predicting density and color for 3D points and pretrained text-to-image diffusion models for providing 2D self-supervision. Specifically, we leverage the SMPL model to provide shape and pose guidance for the generation. We introduce a dual-observation-space design that involves the joint optimization of a canonical space and a posed space that are related by a learnable deformation field. This facilitates the generation of more complete textures and geometry faithful to the target pose. We also jointly optimize the losses computed from the full body and from the zoomed-in 3D head to alleviate the common multi-face ``Janus'' problem and improve facial details in the generated avatars. Extensive evaluations demonstrate that DreamAvatar significantly outperforms existing methods, establishing a new state-of-the-art for text-and-shape guided 3D human avatar generation.

Wi-Fi signals, in contrast to cameras, offer privacy protection and occlusion resilience for some practical scenarios such as smart homes, elderly care, and virtual reality. Recent years have seen remarkable progress in the estimation of single-person 2D pose, single-person 3D pose, and multi-person 2D pose. This paper takes a step forward by introducing \system, a pioneering Wi-Fi system that accomplishes multi-person 3D pose estimation. Person-in-WiFi 3D has two main updates. Firstly, it has a greater number of Wi-Fi devices to enhance the capability for capturing spatial reflections from multiple individuals. Secondly, it leverages the Transformer for end-to-end estimation. Compared to its predecessor, Person-in-WiFi 3D is storage-efficient and fast. We deployed a proof-of-concept system in 4mx3.5m areas and collected a dataset of over 97K frames with seven volunteers. Person-in-WiFi 3D attains 3D joint localization errors of 91.7mm (1-person), 108.1mm (2-person), and 125.3mm (3-person), comparable to cameras and millimeter-wave radars.

Monocular 3D human mesh estimation is an ill-posed problem, characterized by inherent ambiguity and occlusion. While recent probabilistic methods propose generating multiple solutions, little attention is paid to obtaining high-quality estimates from them. To address this limitation, we introduce a versatile framework by first leveraging our novel \textbf{HypoNet} to generate multiple hypotheses, followed by employing a meticulously designed scorer, \textbf{ScoreNet}, to evaluate and select high-quality estimates. Our framework formulates the estimation process as a reverse denoising process, where HypoNet produces a diverse set of plausible estimates that effectively align with the image cues. Subsequently, ScoreNet is employed to rigorously evaluate and rank these estimates based on their quality and finally identify superior ones. Experimental results demonstrate that HypoNet outperforms existing state-of-the-art probabilistic methods as a multi-hypothesis mesh estimator. Moreover, the estimates selected by ScoreNet significantly outperform random generation or simple averaging. Notably, the trained ScoreNet exhibits generalizability, as it can effectively score existing methods and significantly reduce their errors by more than 16mm. Code and models will be publicly released.

This paper tackles the problem of creating relightable and animatable neural avatars from sparse-view (or monocular) videos of dynamic humans under unknown illumination. Previous neural human reconstruction methods produce animatable avatars from sparse views using deformed Signed Distance Fields (SDF) but are non-relightable. While differentiable inverse rendering methods have succeeded in the material recovery of static objects, it is not straightforward to extend them to dynamic humans since it is computationally intensive to compute pixel-surface intersection and light visibility on deformed SDFs for relighting. To solve this challenge, we propose a Hierarchical Distance Query (HDQ) algorithm to approximate the world space SDF under arbitrary human poses. Specifically, we estimate coarse SDF based on a parametric human model and compute fine SDF by exploiting the invariance of SDF w.r.t. local deformation. Based on HDQ, we leverage sphere tracing to efficiently estimate the surface intersection and light visibility. This allows us to develop the first system to recover relightable and animatable neural avatars from sparse or monocular inputs. Experiments show that our approach produces superior results compared to state-of-the-art methods. Our project page is available at https://zju3dv.github.io/relightable_avatar.

The fidelity of relighting is bounded by both geometry and appearance representations. For geometry, both mesh and volumetric approaches have difficulty modeling intricate structures like 3D hair geometry. For appearance, existing relighting models are limited in fidelity and often too slow to render in real-time with high-resolution continuous environments. In this work, we present Relightable Gaussian Codec Avatars, a method to build high-fidelity relightable head avatars that can be animated to generate novel expressions. Our geometry model based on 3D Gaussians can capture 3D-consistent sub-millimeter details such as hair strands and pores on dynamic face sequences. To support diverse materials of human heads such as the eyes, skin, and hair in a unified manner, we present a novel relightable appearance model based on learnable radiance transfer. Together with global illumination-aware spherical harmonics for the diffuse components, we achieve real-time relighting with spatially all-frequency reflections using spherical Gaussians. This appearance model can be efficiently relit under both point light and continuous illumination. We further improve the fidelity of eye reflections and enable explicit gaze control by introducing relightable explicit eye models. Our method outperforms existing approaches without compromising real-time performance. We also demonstrate real-time relighting of avatars on a tethered consumer VR headset, showcasing the efficiency and fidelity of our avatars.

We present a framework for generating full-bodied photorealistic avatars that gesture according to the conversational dynamics of a dyadic interaction. Given speech audio, we output multiple possibilities of gestural motion for an individual, including face, body, and hands. The key behind our method is in combining the benefits of sample diversity from vector quantization with the high-frequency details obtained through diffusion to generate more dynamic, expressive motion. We visualize the generated motion using highly photorealistic avatars that can express crucial nuances in gestures (e.g. sneers and smirks). To facilitate this line of research, we introduce a first-of-its-kind multi-view conversational dataset that allows for photorealistic reconstruction. Experiments show our model generates appropriate and diverse gestures, outperforming both diffusion- and VQ-only methods. Furthermore, our perceptual evaluation highlights the importance of photorealism (vs. meshes) in accurately assessing subtle motion details in conversational gestures. Code and dataset will be publicly released.

Existing multi-person human reconstruction approaches mainly focus on recovering accurate poses or avoiding penetration, but overlook the modeling of close interactions. In this work, we tackle the task of reconstructing closely interactive humans from a monocular video. The main challenge of this task comes from insufficient visual information caused by depth ambiguity and severe inter-person occlusion. In view of this, we propose to leverage knowledge from proxemic behavior and physics to compensate the lack of visual information. This is based on the observation that human interaction has specific patterns following the social proxemics. Specifically, we first design a latent representation based on Vector Quantised-Variational AutoEncoder (VQ-VAE) to model human interaction. A proxemics and physics guided diffusion model is then introduced to denoise the initial distribution. We design the diffusion model as dual branch with each branch representing one individual such that the interaction can be modeled via cross attention. With the learned priors of VQ-VAE and physical constraint as the additional information, our proposed approach is capable of estimating accurate poses that are also proxemics and physics plausible. Experimental results on Hi4D, 3DPW, and CHI3D demonstrate that our method outperforms existing approaches. The code will be made publicly available.

By leveraging temporal dependency in video sequences, multi-frame human pose estimation algorithms have demonstrated remarkable results in complicated situations, such as occlusion, motion blur, and video defocus. These algorithms are predominantly based on heatmaps, resulting in high computation and storage requirements per frame, which limits their flexibility and real-time application in video scenarios, particularly on edge devices. In this paper, we develop an efficient and effective video-based human pose regression method, which bypasses intermediate representations such as heatmaps and instead directly maps the input to the output joint coordinates. Despite the inherent spatial correlation among adjacent joints of the human pose, the temporal trajectory of each individual joint exhibits relative independence. In light of this, we propose a novel Decoupled Space-Time Aggregation network (DSTA) to separately capture the spatial contexts between adjacent joints and the temporal cues of each individual joint, thereby avoiding the conflation of spatiotemporal dimensions. Concretely, DSTA learns a dedicated feature token for each joint to facilitate the modeling of their spatiotemporal dependencies. With the proposed joint-wise local-awareness attention mechanism, our method is capable of efficiently and flexibly utilizing the spatial dependency of adjacent joints and the temporal dependency of each joint itself. Extensive experiments demonstrate the superiority of our method. Notably, our approach is the first regression-based method for multi-frame human pose estimation. Compared to previous regression-based single-frame human pose estimation methods, DSTA significantly enhances performance, achieving an 8.9 mAP improvement on PoseTrack2017. Furthermore, our approach either surpasses or is on par with the state-of-the-art heatmap-based multi-frame human pose estimation methods.

Generalizable face anti-spoofing (FAS) approaches have drawn growing attention due to their robustness for diverse presentation attacks in unseen scenarios. Most previous methods always utilize domain generalization (DG) frameworks via directly aligning diverse source samples into a common feature space.However, these methods neglect the hierarchical relations in FAS samples which may hinder the generalization ability by direct alignment. To address these issues, we propose a novel Hierarchical Prototype-guided Distribution Refinement (HPDR) framework to learn embedding in hyperbolic space, which facilitates the hierarchical relation construction. We also collaborate with prototype learning for hierarchical distribution refinement in hyperbolic space. In detail, we propose the Hierarchical Prototype Learning to simultaneously guide domain alignment and improve the discriminative ability via constraining the multi-level relations between prototypes and instances in hyperbolic space.Moreover, we design a Prototype-oriented Classifier, which further considers relations between the sample and prototypes to improve the robustness of the final decision. Extensive experiments and visualizations demonstrate the effectiveness of our method against previous competitors.

In the academic field, the research on human motion prediction tasks mainly focuses on exploiting the observed information to forecast human movements accurately in the near future horizon. However, a significant gap appears when it comes to the application field, as current models are all trained offline, with fixed parameters that are inherently suboptimal to handle the complex yet ever-changing nature of human behaviors. To bridge this gap, in this paper, we introduce the task of online meta adaptation for human motion prediction, based on the insight that finding "smart weights" capable of swift adjustments to suit different motion contexts along the time is a key to improving predictive accuracy. We propose MoML, which ingeniously borrows the bilevel optimization spirit of model-agnostic meta-learning, to transform previous predictive mistakes into strong inductive biases to guide online adaptation. This is achieved by our MoAdapter blocks that can learn error information by facilitating efficient adaptation via a few gradient steps, which fine-tunes our meta-learned "smart" initialization produced by the generic predictor. Considering real-time requirements in practice, we further propose Fast-MoML, a more efficient variant of MoML that features a closed-form solution instead of conventional gradient update. Experimental results show that our approach can effectively bring many existing offline motion prediction models online, and improves their predictive accuracy.

2D keypoints are commonly used as an additional cue to refine estimated 3D human meshes. Current methods optimize the pose and shape parameters with a reprojection loss on the provided 2D keypoints. Such an approach, while simple and intuitive, has limited effectiveness because the optimal solution is hard to find in ambiguous parameter space and may sacrifice depth. Additionally, divergent gradients from distal joints complicate and deviate the refinement of proximal joints in the kinematic chain. To address these, we introduce Kinematic-Tree Rotation (KITRO), a novel mesh refinement strategy that explicitly models depth and human kinematic-tree structure. KITRO treats refinement from a bone-wise perspective. Unlike previous methods which perform gradient-based optimizations, our method calculates bone directions in closed form. By accounting for the 2D pose, bone length, and parent joint's depth, the calculation results in two possible directions for each child joint. We then use a decision tree to trace binary choices for all bones along the human skeleton's kinematic-tree to select the most probable hypothesis. Our experiments across various datasets and baseline models demonstrate that KITRO significantly improves 3D joint estimation accuracy and achieves an ideal 2D fit simultaneously. Our code available at: https://github.com/MartaYang/KITRO.

In this paper, we present a method to reconstruct the world and multiple dynamic humans in 3D from a monocular video input. As a key idea, we represent both the world and multiple humans via the recently emerging 3D Gaussian Splatting (3D-GS) representation, enabling us to conveniently and efficiently compose and render them together. In particular, we address the scenarios with severely limited and sparse observations in 3D human reconstruction, a common challenge encountered in the real world. To tackle this challenge, we introduce a novel approach to optimize the 3D-GS representation in a canonical space by fusing the sparse cues in the common space, where we leverage a pre-trained 2D diffusion model to $\textbf{synthesize}$ unseen views while keeping the consistency with the observed 2D appearances. We demonstrate our method can reconstruct high-quality animatable 3D humans in various challenging examples, in the presence of occlusion, image crops, few-shot, and extremely sparse observations. The output model provides not only the ability to render the scene in any novel views at arbitrary time instances but also to edit and alter the 3D scene by removing individuals or applying different motions for 3D humans. Once optimization is done, our method is capable of not only rendering the scene in any novel views at arbitrary time instances, but also editing the 3D scene by removing individual humans or applying different motions for each human.Through various experiments, we demonstrate the quality and efficiency of our methods over alternative existing approaches.

We present, PEGASUS, a method for constructing personalized generative 3D face avatars from monocular video sources. As a compositional generative model, our model enables disentangled controls to selectively alter the facial attributes (e.g., hair or nose) of the target individual, while preserving the identity. We present two key approaches to achieve this goal. First, we present a method to construct a person-specific generative 3D avatar by building a synthetic video collection of the target identity with varying facial attributes, where the videos are synthesized by borrowing parts from diverse individuals from other monocular videos. Through several experiments, we demonstrate the superior performance of our approach by generating unseen attributes with high realism. Subsequently, we introduce a zero-shot approach to achieve the same generative modeling more efficiently by leveraging a previously constructed personalized generative model.

The field of 3D detailed human mesh reconstruction has made significant progress in recent years. However, current methods still face challenges when used in industrial applications due to unstable results, low-quality meshes, and a lack of UV unwrapping and skinning weights. In this paper, we present SHERT, a novel pipeline that can reconstruct semantic human meshes with textures and high-precision details. SHERT applies semantic and normal-based sampling between the detailed surface (e.g. mesh and SDF) and the corresponding SMPL-X model to obtain a partially sampled semantic mesh and then generates the complete semantic mesh by our specifically designed self-supervised completion and refinement networks. Using the complete semantic mesh as a basis, we employ a texture diffusion model to create human textures that are driven by both images and text. Our reconstructed meshes have stable UV unwrapping, high-quality triangle meshes, and consistent semantic information. The given SMPL-X model provides semantic information and shape priors, allowing SHERT to perform well even with incorrect and incomplete inputs. The semantic information also makes it easy to substitute and animate different body parts such as the face, body, and hands. Quantitative and qualitative experiments demonstrate that SHERT is capable of producing high-fidelity and robust semantic meshes that outperform state-of-the-art methods.

Recently, transformer-based methods have achieved state-of-the-art prediction quality on human pose estimation(HPE). Nonetheless, most of these top-performing transformer-based models are too computation-consuming and storage-demanding to deploy on edge computing platforms. Those transformer-based models that require fewer resources are prone to under-fitting due to their smaller scale and thus perform notably worse than their larger counterparts. Given this conundrum, we introduce SDPose, a new self-distillation method for improving the performance of small transformer-based models. To mitigate the problem of under-fitting, we design a transformer module named Multi-Cycled Transformer(MCT) based on multiple-cycled forwards to more fully exploit the potential of small model parameters. Further, in order to prevent the additional inference compute-consuming brought by MCT, we introduce a self-distillation scheme, extracting the knowledge from the MCT module to a naive forward model. Specifically, on the MSCOCO validation dataset, SDPose-T obtains 69.7\% mAP with 4.4M parameters and 1.8 GFLOPs. Furthermore, SDPose-S-V2 obtains 73.5\% mAP on the MSCOCO validation dataset with 6.2M parameters and 4.7 GFLOPs, achieving a new state-of-the-art among predominant tiny neural network methods.

We present a lightweight and affordable motion capture method based on two smartwatches and a head-mounted camera. In contrast to the existing approaches that use six or more expert-level IMU devices, our approach is much more cost-effective and convenient. Our method can make wearable motion capture accessible to everyone, enabling 3D full-body motion capture in diverse environments. As a key idea to overcome the extreme sparsity and ambiguities of sensor inputs, we integrate 6D head poses obtained from the head-mounted cameras for motion estimation. To enable capture in expansive indoor and outdoor scenes, we propose an algorithm to track and update floor level changes to define head pose, coupled with a multi-stage Transformer-based regression module. We also introduce novel strategies leveraging visual cues of egocentric images to further enhance the motion capture quality while reducing ambiguities. We demonstrate the performance of our method on various challenging scenarios, including complex outdoor environments and everyday motions including object interactions and social interactions among multiple individuals.

The recovery of occluded human meshes poses challenges for current methods due to the difficulty in extracting effective image features under severe occlusion. In this paper, we introduce DPMesh, an innovative framework for occluded human mesh recovery that capitalizes on the profound knowledge about object structure and spatial relationships embedded in a pre-trained text-to-image diffusion model. Unlike previous methods reliant on conventional backbones for vanilla feature extraction, DPMesh seamlessly integrates the pre-trained denoising U-Net with potent priors as its image backbone and performs a single-step inference to provide occlusion-aware information. To enhance the perception capability for occluded poses, DPMesh incorporates judicious guidance via condition injection, which produces effective controls from 2D observations for the denoising U-Net. Furthermore, we explore a dedicated noisy key-point reasoning approach to mitigate disturbances arising from occlusion and crowded scenarios. This strategy fully unleashes the perceptual capability of the diffusion prior, thereby enhancing accuracy. Extensive quantitative and qualitative experiments affirm the efficacy of our framework, as we outperform state-of-the-art methods on both occlusion-specific and standard datasets, underscoring its ability to achieve precise and robust 3D human mesh recovery, particularly in challenging scenarios involving occlusion and crowded scenes. Code is available at https://github.com/EternalEvan/DPMesh.

We introduce Diffusion Parametric Head Models (DPHMs), a generative model that enables robust volumetric head reconstruction and tracking from monocular depth sequences. While recent volumetric head models, such as NPHMs, can now excel in representing high-fidelity head geometries, tracking and reconstructing heads from real-world single-view depth sequences remains very challenging, as the fitting to partial and noisy observations is underconstrained. To tackle these challenges, we propose a latent diffusion-based prior to regularize volumetric head reconstruction and tracking. This prior-based regularizer effectively constrains the identity and expression codes to lie on the underlying latent manifold which represents plausible head shapes. To evaluate the effectiveness of the diffusion-based prior, we collect a dataset of monocular Kinect sequences consisting of various complex facial expression motions and rapid transitions. We compare our method to state-of-the-art tracking methods and demonstrate improved head identity reconstruction as well as robust expression tracking.

This paper presents a novel Kinematics and Trajectory Prior Knowledge-Enhanced Transformer (KTPFormer), which overcomes the weakness in existing transformer-based methods for 3D human pose estimation that the derivation of Q, K, V vectors in their self-attention mechanisms are all based on simple linear mapping. We propose two prior attention modules, namely Kinematics Prior Attention (KPA) and Trajectory Prior Attention (TPA) to take advantage of the known anatomical structure of the human body and motion trajectory information, to facilitate effective learning of global dependencies and features in the multi-head self-attention. KPA models kinematic relationships in the human body by constructing a topology of kinematics, while TPA builds a trajectory topology to learn the information of joint motion trajectory across frames. Yielding Q, K, V vectors with prior knowledge, the two modules enable KTPFormer to model both spatial and temporal correlations simultaneously. Extensive experiments on three benchmarks (Human3.6M, MPI-INF-3DHP and HumanEva) show that KTPFormer achieves superior performance in comparison to state-of-the-art methods. More importantly, our KPA and TPA modules have lightweight plug-and-play designs and can be integrated into various transformer-based networks (i.e., diffusion-based) to improve the performance with only a very small increase in the computational overhead. The code is available at: https://github.com/JihuaPeng/KTPFormer.

This paper presents a new approach for the detection of fake videos, based on the analysis of style latent vectors and their abnormal behavior in temporal changes in the generated videos. We discovered that the generated facial videos suffer from the temporal distinctiveness in the temporal changes of style latent vectors, which are inevitable during the generation of temporally stable videos with various facial expressions and geometric transformations. Our framework utilizes the Style GRU module, trained by contrastive learning, to represent the dynamic properties of style latent vectors. Additionally, we introduce a style attention module that integrates Style GRU-generated features with content-based features, enabling the detection of visual and temporal artifacts. We demonstrate our approach across various benchmark scenarios in deepfake detection, showing its superiority in cross-dataset and cross-manipulation scenarios. Through further analysis, we also validate the importance of using temporal changes of style latent vectors to improve the generality of deepfake video detection.

We propose EMAGE, a framework to generate full-body human gestures from audio and masked gestures, encompassing facial, local body, hands, and global movements. To achieve this, we first introduce BEAT2 (BEAT-SMPLX-FLAME), a new mesh-level holistic co-speech dataset. BEAT2 combines a MoShed SMPL-X body with FLAME head parameters and further refines the modeling of head, neck, and finger movements, offering a community-standardized, high-quality 3D motion captured dataset. EMAGE leverages masked body gesture priors during training to boost inference performance. It involves a Masked Audio Gesture Transformer, facilitating joint training on audio-to-gesture generation and masked gesture reconstruction to effectively encode audio and body gesture hints. Encoded body hints from masked gestures are then separately employed to generate facial and body movements. Moreover, EMAGE adaptively merges speech features from the audio's rhythm and content and utilizes four compositional VQ-VAEs to enhance the results' fidelity and diversity. Experiments demonstrate that EMAGE generates holistic gestures with state-of-the-art performance and is flexible in accepting predefined spatial-temporal gesture inputs, generating complete, audio-synchronized results. Our code and dataset are available. https://pantomatrix.github.io/EMAGE/

Human-centric Point Cloud Video Understanding (PVU) is an emerging field focused on extracting and interpreting human-related features from sequences of human point clouds, further advancing downstream human-centric tasks and applications. Previous works usually focus on tackling one specific task and rely on huge labeled data, which has poor generalization capability. Considering that human has specific characteristics, including the structural semantics of human body and the dynamics of human motions, we propose a unified framework to make full use of the prior knowledge and explore the inherent features in the data itself for generalized human-centric point cloud video understanding. Extensive experiments demonstrate that our method achieves state-of-the-art performance on various human-related tasks, including action recognition and 3D pose estimation. All datasets and code will be released soon.

Real-time rendering of photorealistic and controllable human avatars stands as a cornerstone in Computer Vision and Graphics. While recent advances in neural implicit rendering have unlocked unprecedented photorealism for digital avatars, real-time performance has mostly been demonstrated for static scenes only. To address this, we propose ASH, an animatable Gaussian splatting approach for photorealistic rendering of dynamic humans in real time. We parameterize the clothed human as animatable 3D Gaussians, which can be efficiently splatted into image space to generate the final rendering. However, naively learning the Gaussian parameters in 3D space poses a severe challenge in terms of compute. Instead, we attach the Gaussians onto a deformable character model, and learn their parameters in 2D texture space, which allows leveraging efficient 2D convolutional architectures that easily scale with the required number of Gaussians. We benchmark ASH with competing methods on pose-controllable avatars, demonstrating that our method outperforms existing real-time methods by a large margin and shows comparable or even better results than offline methods.

Even the best current algorithms for estimating body 3D shape and pose yield results that include body self-intersections. In this paper, we present CLOAF, which exploits the diffeomorphic nature of Ordinary Differential Equations to eliminate such self-intersections while still imposing body shape constraints. We show that, unlike earlier approaches to addressing this issue, ours completely eliminates the self-intersections without compromising the accuracy of the reconstructions. Being differentiable, CLOAF can be used to fine-tune pose and shape estimation baselines to improve their overall performance and eliminate self-intersections in their predictions. Furthermore, we demonstrate how our CLOAF strategy can be applied to practically any motion field induced by the user.CLOAF also makes it possible to edit motion to interact with the environment without worrying about potential collision or loss of body-shape prior.

Monocular egocentric 3D human motion capture is a challenging and actively researched problem. Existing methods use synchronously operating visual sensors (e.g. RGB cameras) and often fail under low lighting and fast motions, which can be restricting in many applications involving head-mounted devices. In response to the existing limitations, this paper 1) introduces a new problem, i.e. 3D human motion capture from an egocentric monocular event camera with a fisheye lens, and 2) proposes the first approach to it called EventEgo3D (EE3D). Event streams have high temporal resolution and provide reliable cues for 3D human motion capture under high-speed human motions and rapidly changing illumination. The proposed EE3D framework is specifically tailored for learning with event streams in the LNES representation, enabling high 3D reconstruction accuracy. We also design a prototype of a mobile head-mounted device with an event camera and record a real dataset with event observations and the ground-truth 3D human poses (in addition to the synthetic dataset). Our EE3D demonstrates robustness and superior 3D accuracy compared to existing solutions across various challenging experiments while supporting real-time 3D pose update rates of 140Hz. We will release the spec ifications of our device, the datasets and the source codes.

Comparing different age estimation methods poses a challenge due to the unreliability of published results stemming from inconsistencies in the benchmarking process. Previous studies have reported continuous performance improvements over the past decade using specialized methods; however, our findings challenge these claims. This paper identifies two trivial, yet persistent issues with the currently used evaluation protocol and describes how to resolve them. We offer an extensive comparative analysis for state-of-the-art facial age estimation methods. Surprisingly, we find that the performance differences between the methods are negligible compared to the effect of other factors, such as facial alignment, facial coverage, image resolution, model architecture, or the amount of data used for pretraining. We use the gained insights to propose using FaRL as the backbone model and demonstrate its effectiveness on all public datasets. We make the source code and exact data splits public on GitHub and in the supplementary material.

We present the first approach to render highly realistic free-viewpoint videos of a human actor in general apparel, from sparse multi-view recording to display, in real-time at an unprecedented 4K resolution. At inference, our method only requires four camera views of the moving actor and the respective 3D skeletal pose. It handles actors in wide clothing, and reproduces even fine-scale dynamic detail, e.g. clothing wrinkles, face expressions, and hand gestures. At training time, our learning-based approach expects dense multi-view video and a rigged static surface scan of the actor. Our method comprises three main stages. Stage 1 is a skeleton-driven neural approach for high-quality capture of the detailed dynamic mesh geometry. Stage 2 is a novel solution to create a view-dependent texture using four testtime camera views as input. Finally, stage 3 comprises a new image-based refinement network rendering the final 4K image given the output from the previous stages. Our approach establishes a new benchmark for real-time rendering resolution and quality using sparse input camera views, unlocking possibilities for immersive telepresence. Code and data is available on our project page.

Remarkable strides have been made in reconstructing static scenes or human bodies from monocular videos. Yet, the two problems have largely been approached independently, without much synergy. Most visual SLAM methods can only reconstruct camera trajectories and scene structures up to scale, while most HMR methods reconstruct human meshes in metric scale but fall short in reasoning with cameras and scenes. This work introduces Synergistic Camera and Human Reconstruction (SynCHMR) to marry the best of both worlds. Specifically, we design Human-aware Metric SLAM to reconstruct metric-scale camera poses and scene point clouds using camera-frame HMR as a strong prior, addressing depth, scale, and dynamic ambiguities. Conditioning on the dense scene recovered, we further learn a Scene-aware SMPL Denoiser to enhance world-frame HMR by incorporating spatio-temporal coherency and dynamic scene constraints. Together, they lead to consistent reconstructions of camera trajectories, human meshes, and dense scene point clouds in a common world frame.

Improving the 3D facial fidelity and avoiding the uncanny valley effect is critically dependent on accurate 3D facial performance capture. Due to the widespread availability of 2D videos, recent methods focus on monocular 3D face tracking. However, these methods often fall short in capturing precise facial movements due to limitations in their network architecture, training, and evaluation processes. Addressing these challenges, we propose a novel face tracker, FlowFace, that introduces an innovative 2D alignment network for dense per-vertex alignment. Unlike prior work, FlowFace is trained on high-quality 3D scan annotations rather than weak supervision or synthetic data. Our 3D model fitting module jointly fits a 3D face model from one or many observations, integrating existing neutral shape priors for enhanced identity and expression disentanglement and per-vertex deformations for detailed facial feature reconstruction. Additionally, we propose a novel metric and benchmark for assessing tracking accuracy. Our method exhibits superior performance on both custom and publicly available benchmarks. We further validate the effectiveness of our tracker by generating high-quality data, which leads to performance gains on downstream tasks.

Recent advances in 3D avatar generation have gained significant attentions. These breakthroughs aim to produce more realistic animatable avatars, narrowing the gap between virtual and real-world experiences. Most of existing works employ Score Distillation Sampling (SDS) loss, combined with a differentiable renderer and text condition, to guide a diffusion model in generating 3D avatars. However, SDS often generates oversmoothed results with few facial details, thereby lacking the diversity compared with ancestral sampling. On the other hand, other works generate 3D avatar from a single image, where the challenges of unwanted lighting effects, perspective views, and inferior image quality make them difficult to reliably reconstruct the 3D face meshes with the aligned complete textures. In this paper, we propose a novel 3D avatar generation approach termed UltrAvatar with enhanced fidelity of geometry, and superior quality of physically based rendering (PBR) textures without unwanted lighting. To this end, the proposed approach presents a diffuse color extraction model and an authenticity guided texture diffusion model. The former removes the unwanted lighting effects to reveal true diffuse colors so that the generated avatars can be rendered under various lighting conditions. The latter follows two gradient-based guidances for generating PBR textures to render diverse face-identity features and details better aligning with 3D mesh geometry. We demonstrate the effectiveness and robustness of the proposed method, outperforming the state-of-the-art methods by a large margin in the experiments.

Our paper aims to generate diverse and realistic animal motion sequences from textual descriptions, without a large-scale animal text-motion dataset. While the task of text-driven human motion synthesis is already extensively studied and benchmarked, it remains challenging to transfer this success to other skeleton structures with limited data. In this work, we design a model architecture that imitates Generative Pretraining Transformer (GPT), utilizing prior knowledge learned from human data to the animal domain. We jointly train motion autoencoders for both animal and human motions and at the same time optimize through the similarity scores among human motion encoding, animal motion encoding, and text CLIP embedding. Presenting the first solution to this problem, we are able to generate animal motions with high diversity and fidelity, quantitatively and qualitatively outperforming the results of training human motion generation baselines on animal data. Additionally, we introduce AnimalML3D, the first text-animal motion dataset with 1240 animation sequences spanning 36 different animal identities. We hope this dataset would mediate the data scarcity problem in text-driven animal motion generation, providing a new playground for the research community.

Text-guided 3D face synthesis has achieved remarkable results by leveraging text-to-image (T2I) diffusion models. However, most existing works focus solely on the direct generation, ignoring the editing of the faces, restricting them from synthesizing customized 3D faces through iterative editing. In this paper, we propose a unified text-guided framework from face generation to editing. In the generation stage, we propose a geometry-texture decoupled generation to mitigate the loss of geometric details caused by coupling. Besides, decoupling enables us to utilize the generated geometry as a condition for texture generation, yielding highly geometry-texture aligned results. We further employ a fine-tuned texture diffusion model to enhance texture quality in both RGB and YUV space. In the editing stage, we first employ a pre-trained diffusion model to update facial geometry or texture based on the texts. To enable sequential editing, we introduce a UV domain consistency preservation regularization, preventing unintentional changes to irrelevant facial attributes. Besides, we propose a self-guided consistency weight strategy to improve editing efficacy while preserving consistency. Through comprehensive experiments and comparisons with existing methods, we showcase our method's superiority in face synthesis.

Human facial action units (AUs) are mutually related in a hierarchical manner, as not only they are associated with each other in both spatial and temporal domains but also AUs located in the same/close facial regions show stronger relationships than those of different facial regions. While none of existing approach thoroughly model such hierarchical inter-dependencies among AUs, this paper proposes to comprehensively model multi-scale AU-related dynamic and hierarchical spatio-temporal relationship among AUs for their occurrences recognition. Specifically, we first propose a novel multi-scale temporal differencing network with an adaptive weighting block to explicitly capture facial dynamics across frames at different spatial scales, which specifically considers the heterogeneity of range and magnitude in different AUs' activation. Then, a two-stage strategy is introduced to hierarchically model the relationship among AUs based on their spatial distribution (i.e., local and cross-region AU relationship modelling). Experimental results achieved on BP4D and DISFA show that our approach is the new state-of-the-art in the field of AU occurrence recognition. Our code is publicly available at https://github.com/CVI-SZU/MDHR.

For human-centric large-scale scenes, fine-grained modeling for 3D human global pose and shape is significant for scene understanding and can benefit many real-world applications. In this paper, we present LiveHPS, a novel single LiDAR-based approach for scene-level Human Pose and Shape estimation without any limitation of light conditions and wearable devices. In particular, we design a distillation mechanism to mitigate the distribution-varying effect of LiDAR point clouds and exploit the temporal-spatial geometric and dynamic information existing in consecutive frames to solve the occlusion and noise disturbance. LiveHPS, with its efficient configuration and high-quality output, is well-suited for real-world applications.Moreover, we propose a huge human motion dataset, named FreeMotion, which is collected in various scenarios with diverse human poses, shapes and translations. It consists of multi-modal and multi-view acquisition data from calibrated and synchronized LiDARs, cameras, and IMUs. Extensive experiments on our new dataset and other public datasets demonstrate the SOTA performance and robustness of our approach. We will release our code and dataset soon.

In this paper, we abstract the process of people hearing speech, extracting meaningful cues, and creating various dynamically audio-consistent talking faces, termed Listening and Imagining, into the task of high-fidelity diverse talking faces generation from a single audio. Specifically, it involves two critical challenges: one is to effectively decouple identity, content, and emotion from entangled audio, and the other is to maintain intra-video diversity and inter-video consistency. To tackle the issues, we first dig out the intricate relationships among facial factors and simplify the decoupling process, tailoring a Progressive Audio Disentanglement for accurate facial geometry and semantics learning, where each stage incorporates a customized training module responsible for a specific factor. Secondly, to achieve visually diverse and audio-synchronized animation solely from input audio within a single model, we introduce the Controllable Coherent Frame generation, which involves the flexible integration of three trainable adapters with frozen Latent Diffusion Models (LDMs) to focus on maintaining facial geometry and semantics, as well as texture and temporal coherence between frames. In this way, we inherit high-quality diverse generation from LDMs while significantly improving their controllability at a low training cost. Extensive experiments demonstrate the flexibility and effectiveness of our method in handling this paradigm. The codes will be released at https://github.com/modelscope/facechain.

Depression Recognition (DR) poses a considerable challenge,especially in the context of the growing concerns surroundingprivacy. Traditional automatic diagnosis of DRtechnology necessitates the use of facial images, undoubtedlyexpose the patient identity features and poses privacyrisks. In order to mitigate the potential risks associatedwith the inappropriate disclosure of patient facial images,we design a new imaging system to erase the identity informationof captured facial images while retain diseaserelevantfeatures. It is irreversible for identity informationrecovery while preserving essential disease-related characteristicsnecessary for accurate DR. More specifically,we try to record a de-identified facial image (erasing theidentifiable features as much as possible) by a learnablelens, which is optimized in conjunction with the followingDR task as well as a range of face analysis related auxiliarytasks in an end-to-end manner. These aforementionedstrategies form our final Optical deep Depression Recognitionnetwork (OpticalDR). Experiments on CelebA, AVEC2013, and AVEC 2014 datasets demonstrate that our OpticalDRhas achieved state-of-the-art privacy protection performancewith an average AUC of 0.51 on popular facialrecognition models, and competitive results for DR withMAE/RMSE of 7.53/8.48 on AVEC 2013 and 7.89/8.82 onAVEC 2014, respectively.

Self-supervised landmark estimation is a challenging task that demands the formation of locally distinct feature representations to identify sparse facial landmarks in the absence of annotated data. To tackle this task, existing state-of-the-art (SOTA) methods (1) extract coarse features from backbones that are trained with instance-level self-supervised learning (SSL) paradigms, which neglect the dense prediction nature of the task, (2) aggregate them into memory-intensive hypercolumn formations, and (3) supervise lightweight projector networks to naively establish full local correspondences among all pairs of spatial features. In this paper, we introduce SCE-MAE, a framework that (1) leverages the MAE, a region-level SSL method that naturally better suits the landmark prediction task, (2) operates on the vanilla feature map instead of on expensive hypercolumns, and (3) employs a Correspondence Approximation and Refinement Block (CARB) that utilizes a simple density peak clustering algorithm and our proposed Locality-Constrained Repellence Loss to directly hone only select local correspondences. We demonstrate through extensive experiments that SCE-MAE is highly effective and robust, outperforming existing SOTA methods by large margins of $\sim$20\%-44\% on the landmark matching and $\sim$9\%-15\% on the landmark detection tasks.

We address the problem of regressing 3D human pose and shape from a single image, with a focus on 3D accuracy. The current best methods leverage large datasets of 3D pseudo-ground-truth (p-GT) and 2D keypoints, leading to robust performance. With such methods, however, we observe a paradoxical decline in 3D pose accuracy with increasing 2D accuracy. This is caused by biases in the p-GT and the use of an approximate camera projection model. We quantify the error induced by current camera models and show that fitting 2D keypoints and p-GT accurately causes incorrect 3D poses. Our analysis defines the invalid distances within which minimizing 2D and p-GT losses is detrimental. We use this to formulate a new loss, “Threshold-Adaptive Loss Scaling” (TALS), that penalizes gross 2D and p-GT errors but not smaller ones. With such a loss, there are many 3D poses that could equally explain the 2D evidence. To reduce this ambiguity we need a prior over valid human poses but such priors can introduce unwanted bias. To address this, we exploit a tokenized representation of human pose and reformulate the problem as token prediction. This restricts the estimated poses to the space of valid poses, effectively improving robustness to occlusion. Extensive experiments on the EMDB and 3DPW datasets show that our reformulated loss and tokenization allows us to train on in-the-wild data while improving 3D accuracy over the state-of-the-art. Our models and code are available for research at https://tokenhmr.is.tue.mpg.de.

We propose Diffusion Noise Optimization (DNO), a new method that effectively leverages existing motion diffusion models as motion priors for a wide range of motion-related tasks. Instead of training a task-specific diffusion model for each new task, DNO operates by optimizing the diffusion latent noise of an existing pre-trained text-to-motion model. Given the corresponding latent noise of a human motion, it propagates the gradient from the target criteria defined on the motion space through the whole denoising process to update the diffusion latent noise. As a result, DNO supports any use cases where criteria can be defined as a function of motion. In particular, we show that, for motion editing and control, DNO outperforms existing methods in both achieving the objective and preserving the motion content. DNO accommodates a diverse range of editing modes, including changing trajectory, pose, joint locations, or avoiding newly added obstacles. In addition, DNO is effective in motion denoising and completion, producing smooth and realistic motion from noisy and partial inputs. DNO achieves these results at inference time without the need for model retraining, offering great versatility for any defined reward or loss function on the motion representation.

We present M&M VTO–a mix and match virtual try-on method that takes as input multiple garment images, text description for garment layout and an image of a person. An example input includes: an image of a shirt, an image of a pair of pants, "rolled sleeves, shirt tucked in", and an image of a person. The output is a visualization of how those garments (in the desired layout) would look like on the given person. Key contributions of our method are: 1) a single stage diffusion based model, with no super resolution cascading, that allows to mix and match multiple garments at $1024\mathord\times\mathord512$ resolution preserving and warping intricate garment details, 2) architecture design (VTO UNet Diffusion Transformer) to disentangle denoising from person specific features, allowing for a highly effective finetuning strategy for identity preservation (6MB model per individual vs 4GB achieved with, e.g., dreambooth finetuning); solving a common identity loss problem in current virtual try-on methods, 3) layout control for multiple garments via text inputs specifically finetuned over PaLI-3 for virtual try-on task. Experimental results indicate that M&M VTO achieves state-of-the-art performance both qualitatively and quantitatively, as well as opens up new opportunities for virtual try-on via language-guided and multi-garment try-on.

Large Language Models(LLMs) have shown remarkable emergent abilities in unifying almost all (if not every) NLP tasks. In the human motion-related realm, however, researchers still develop siloed models for each task. Inspired by InstuctGPT and the generalist concept behind Gato, we introduce \textbf{AvatarGPT}, an All-in-One framework for motion understanding, planning, generations as well as other tasks such as motion in-between synthesis. AvatarGPT treats each task as one type of instruction fine-tuned on the shared LLM. All the tasks are seamlessly interconnected with language as the universal interface, constituting a closed-loop within the framework. To achieve this, human motion sequences are first encoded as discrete tokens, which serve as the extended vocabulary of LLM. Then, an unsupervised pipeline to generate natural language descriptions of human action sequences from in-the-wild videos is developed. Finally, all tasks are jointly trained. Extensive experiments show that AvatarGPT achieves SOTA on low-level tasks, and promising results on high-level tasks, demonstrating the effectiveness of our proposed All-in-One framework. Moreover, for the first time, AvatarGPT enables a principled approach by iterative traversal of the tasks within the closed-loop for unlimited long-motion synthesis.

3D hand pose estimation has found broad application in areas such as gesture recognition and human-machine interaction tasks. As performance improves, the complexity of the systems also increases, which can limit the comparative analysis and practical implementation of these methods. In this paper, we propose a simple yet effective baseline that not only surpasses state-of-the-art (SOTA) methods but also demonstrates computational efficiency. To establish this baseline, we abstract existing work into two components: a token generator and a mesh regressor, and then examine their core structures. A core structure, in this context, is one that fulfills intrinsic functions, brings about significant improvements, and achieves excellent performance without unnecessary complexities. Our proposed approach is decoupled from any modifications to the backbone, making it adaptable to any modern models. Our method outperforms existing solutions, achieving state-of-the-art (SOTA) results across multiple datasets. On the FreiHAND dataset, our approach produced a PA-MPJPE of 5.7mm and a PA-MPVPE of 6.0mm. Similarly, on the Dexycb dataset, we observed a PA-MPJPE of 5.5mm and a PA-MPVPE of 5.0mm. As for performance speed, our method reached up to 33 frames per second (fps) when using HRNet and up to 70 fps when employing FastViT-MA36

Rigging and skinning clothed human avatars is a challenging task and traditionally requires a lot of manual work and expertise. Recent methods addressing it either generalize across different characters or focus on capturing the dynamics of a single character observed under different pose configurations. However, the former methods typically predict solely static skinning weights, which perform poorly for highly articulated poses, and the latter ones either require dense 3D character scans in different poses or cannot generate an explicit mesh with vertex correspondence over time. To address these challenges, we propose a fully automated approach for creating a fully rigged character with pose-dependent skinning weights, which can be solely learned from multi-view video. Therefore, we first acquire a rigged template, which is then statically skinned. Next, a coordinate-based MLP learns a skinning weights field parameterized over the position in a canonical pose space and the respective pose. Moreover, we introduce our pose- and view-dependent appearance field allowing us to differentiably render and supervise the posed mesh using multi-view imagery. We show that our approach outperforms state-of-the-art while not relying on dense 4D scans. More details can be found on our project page.

Gestures play a key role in human communication. Recent methods for co-speech gesture generation, while managing to generate beat-aligned motions, struggle generating gestures that are semantically aligned with the utterance. Compared to beat gestures that align naturally to the audio signal, semantically coherent gestures require modeling the complex interactions between the language and human motion, and can be controlled by focusing on certain words. Therefore, we present ConvoFusion, a diffusion-based approach for multi-modal gesture synthesis, which can not only generate gestures based on multi-modal speech inputs, but can also facilitate controllability in gesture synthesis. Our method proposes two guidance objectives that allow the users to modulate the impact of different conditioning modalities (e.g. audio vs text) as well as to choose certain words to be emphasized during gesturing. Our method is versatile in that it can be trained either for generating monologue gestures or even the conversational gestures. To further advance the research on multi-party interactive gestures, the DnD Group Gesture dataset is released, which contains 6 hours of gesture data showing 5 people interacting with one another. We compare our method with several recent works and demonstrate effectiveness of our method on a variety of tasks. We urge the reader to watch our supplementary video at https://vcai.mpi-inf.mpg.de/projects/ConvoFusion/.

Character animation in real-world scenarios necessitates a variety of constraints, such as trajectories, key-frames, interactions, etc. Existing methodologies typically treat single or a finite set of these constraint(s) as separate control tasks. These methods are often specialized, and the tasks they address are rarely extendable or customizable.We categorize these as solutions to the close-set motion control problem. In response to the complexity of practical motion control, we propose and attempt to solve the open-set motion control problem. This problem is characterized by an open and fully customizable set of motion control tasks.To address this, we introduce a new paradigm, programmable motion generation. In this paradigm, any given motion control task is broken down into a combination of atomic constraints. These constraints are then programmed into an error function that quantifies the degree to which a motion sequence adheres to them. We utilize a pre-trained motion generation model and optimize its latent code to minimize the error function of the generated motion.Consequently, the generated motion not only inherits the prior of the generative model but also satisfies the requirements of the compounded constraints.Our experiments demonstrate that our approach can generate high-quality motions when addressing a wide range of unseen tasks. These tasks encompass motion control by motion dynamics, geometric constraints, physical laws, interactions with scenes, objects or the character's own body parts, etc. All of these are achieved in a unified approach, without the need for ad-hoc paired training data collection or specialized network designs.During the programming of novel tasks, we observed the emergence of new skills beyond those of the prior model.With the assistance of large language models, we also achieved automatic programming. We hope that this work will pave the way for the motion control of general AI agents.

Deep-learning-based gaze estimation approaches often suffer from notable performance degradation in unseen target domains. One of the primary reasons is that the Fully Connected layer is highly prone to overfitting when mapping the high-dimensional image feature to 3D gaze. In this paper, we propose Analytical Gaze Generalization framework (AGG) to improve the generalization ability of gaze estimation models without touching target domain data. The AGG consists of two modules, the Geodesic Projection Module (GPM) and the Sphere-Oriented Training (SOT). GPM is a generalizable replacement of FC layer, which projects high-dimensional image features to 3D space analytically to extract the principle components of gaze. Then, we propose Sphere-Oriented Training (SOT) to incorporate the GPM into the training process and further improve cross-domain performances. Experimental results demonstrate that the AGG effectively alleviate the overfitting problem and consistently improves the cross-domain gaze estimation accuracy in 12 cross-domain settings, without requiring any target domain data. The insight from the Analytical Gaze Generalization framework has the potential to benefit other regression tasks with physical meanings.

Annotating gaze is an expensive and time-consuming endeavor, requiring costly eye-trackers or complex geometric calibration procedures. Although some eye-based unsupervised gaze representation learning methods have been proposed, the quality of gaze representation extracted by these methods degrades severely when the head pose is large. In this paper, we present the Multi-View Dual-Encoder (MV-DE), a framework designed to learn gaze representations from unlabeled multi-view face images. Through the proposed Dual-Encoder architecture and the multi-view gaze representation swapping strategy, the MV-DE successfully disentangles gaze from general facial information and derives gaze representations closely tied to the subject's eyeball rotation without gaze label. Experimental results illustrate that the gaze representations learned by the MV-DE can be used in downstream tasks, including gaze estimation and redirection. Gaze estimation results indicates that the proposed MV-DE displays notably higher robustness to uncontrolled head movements when compared to state-of-the-art (SOTA) unsupervised learning methods.

3D human generation is increasingly significant in various applications. However, the direct use of 2D generative methods in 3D generation often results in losing local details, while methods that reconstruct geometry from generated images struggle with global view consistency. In this work, we introduce Joint2Human, a novel method that leverages 2D diffusion models to generate detailed 3D human geometry directly, ensuring both global structure and local details. To achieve this, we employ the Fourier occupancy field (FOF) representation, enabling the direct generation of 3D shapes as preliminary results with 2D generative models. With the proposed high-frequency enhancer and the multi-view recarving strategy, our method can seamlessly integrate the details from different views into a uniform global shape. To better utilize the 3D human prior and enhance control over the generated geometry, we introduce a compact spherical embedding of 3D joints. This allows for an effective guidance of pose during the generation process. Additionally, our method can generate 3D humans guided by textual inputs. Our experimental results demonstrate the capability of our method to ensure global structure, local details, high resolution, and low computational cost simultaneously. More results and the code can be found on our project page at http://cic.tju.edu.cn/faculty/likun/projects/Joint2Human.

We present DiffHuman, a probabilistic method for photorealistic 3D human reconstruction from a single RGB image. Despite the ill-posed nature of this problem, most methods are deterministic and output a single solution, often resulting in a lack of geometric detail and blurriness in unseen or uncertain regions. In contrast, DiffHuman predicts a distribution over 3D reconstructions conditioned on an image, which allows us to sample multiple detailed 3D avatars that are consistent with the input image.DiffHuman is implemented as a conditional diffusion model that denoises partial observations of an underlying pixel-aligned 3D representation. In testing, we can sample a 3D shape by iteratively denoising renderings of the predicted intermediate representation.Further, we introduce an additional generator neural network that approximates rendering with considerably reduced runtime (55x speed up), resulting in a novel dual-branch diffusion framework. We evaluate the effectiveness of our approach through various experiments. Our method can produce diverse, more detailed reconstructions for the parts of the person not observed in the image, and has competitive performance for the surface reconstruction of visible parts.

Current approaches in Group Activity Recognition (GAR) predominantly emphasize Human Relations (HRs) while often neglecting the impact of Human-Object Interactions (HOIs). This study prioritizes the consideration of both HRs and HOIs, emphasizing their interdependence. Notably, employing Granger Causality Tests reveals the presence of bidirectional causality between HRs and HOIs. Leveraging this insight, we propose a Bidirectional-Causal GAR network.This network establishes a causality communication channel while modeling relations and interactions, enabling reciprocal enhancement between human-object interactions and human relations, ensuring their mutual consistency. Additionally, an Interaction Module is devised to effectively capture the dynamic nature of human-object interactions.Comprehensive experiments conducted on two publicly available datasets showcase the superiority of our proposed method over state-of-the-art approaches.

We present HumanNeRF-SE, a simple yet effective method that synthesizes diverse novel pose images with simple input. Previous HumanNeRF works require a large number of optimizable parameters to fit the human images. Instead, we reload these approaches by combining explicit and implicit human representations to design both generalized rigid deformation and specific non-rigid deformation. Our key insight is that explicit shape can reduce the sampling points used to fit implicit representation, and frozen blending weights from SMPL constructing a generalized rigid deformation can effectively avoid overfitting and improve pose generalization performance. Our architecture involving both explicit and implicit representation is simple yet effective. Experiments demonstrate our model can synthesize images under arbitrary poses with few-shot input and increase the speed of synthesizing images by 15 times through a reduction in computational complexity without using any existing acceleration modules. Compared to the state-of-the-art HumanNeRF studies, HumanNeRF-SE achieves better performance with fewer learnable parameters and less training time.

Human pose and shape (HPS) estimation with lensless imaging is not only beneficial to privacy protection but also can be used in covert surveillance scenarios due to the small size and simple structure of this device. However, this task presents significant challenges due to the inherent ambiguity of the captured measurements and lacks effective methods for directly estimating human pose and shape from lensless data. In this paper, we propose the first end-to-end framework to recover 3D human poses and shapes from lensless measurements to our knowledge. We specifically design a multi-scale lensless feature decoder to decode the lensless measurements through the optically encoded mask for efficient feature extraction. We also propose a double-head auxiliary supervision mechanism to improve the estimation accuracy of human limb ends. Besides, we establish a lensless imaging system and verify the effectiveness of our method on various datasets acquired by our lensless imaging system. The code and dataset are available at https://cic.tju.edu.cn/faculty/likun/projects/LPSNet.

This paper studies the human image animation task, which aims to generate a video of a certain reference identity following a particular motion sequence. Existing animation works typically employ the frame-warping technique to animate the reference image towards the target motion. Despite achieving reasonable results, these approaches face challenges in maintaining temporal consistency throughout the animation due to the lack of temporal modeling and poor preservation of reference identity. In this work, we introduce MagicAnimate, a diffusion-based framework that aims at enhancing temporal consistency, preserving reference image faithfully, and improving animation fidelity. To achieve this, we first develop a video diffusion model to encode temporal information. Second, to maintain the appearance coherence across frames, we introduce a novel appearance encoder to retain the intricate details of the reference image. Leveraging these two innovations, we further employ a simple video fusion technique to encourage smooth transitions for long video animation. Empirical results demonstrate the superiority of our method over baseline approaches on two benchmarks. Notably, our approach outperforms the strongest baseline by over 38% in terms of video fidelity on the challenging TikTok dancing dataset. Code and model will be made available.

Real-time multi-person pose estimation presents significant challenges in balancing speed and precision. While two-stage top-down methods slow down as the number of people in the image increases, existing one-stage methods often fail to simultaneously deliver high accuracy and real-time performance. This paper introduces RTMO, a one-stage pose estimation framework that seamlessly integrates coordinate classification by representing keypoints using dual 1-D heatmaps within the YOLO architecture, achieving accuracy comparable to top-down methods while maintaining high speed. We propose a dynamic coordinate classifier and a tailored loss function for heatmap learning, specifically designed to address the incompatibilities between coordinate classification and dense prediction models. RTMO outperforms state-of-the-art one-stage pose estimators, achieving 1.1% higher AP on COCO while operating about 9 times faster with the same backbone. Our largest model, RTMO-l, attains 74.8% AP on COCO val2017 and 141 FPS on a single V100 GPU, demonstrating its efficiency and accuracy. The code and models are available at https://github.com/open-mmlab/mmpose/tree/main/projects/rtmo.

We investigate a new task in human motion prediction, which is predicting motions under unexpected physical perturbation potentially involving multiple people. Compared with existing research, this task involves predicting less controlled, unpremeditated and pure reactive motions in response to external impact and how such motions can propagate through people. It brings new challenges such as data scarcity and predicting complex interactions. To this end, we propose a new method capitalizing differential physics and deep neural networks, leading to an explicit Latent Differential Physics (LDP) model. Through experiments, we demonstrate that LDP has high data efficiency, outstanding prediction accuracy, strong generalizability and good explainability. Since there is no similar research, a comprehensive comparison with 11 adapted baselines from several relevant domains is conducted, showing LDP outperforming existing research both quantitatively and qualitatively, improving prediction accuracy by as much as 70%, and demonstrating significantly stronger generalization.

Human perception and understanding is a major domain of computer vision which, like many other vision subdomains recently, stands to gain from the use of large models pre-trained on large datasets. We hypothesize that the most common pre-training strategy of relying on general purpose, object-centric image datasets such as ImageNet, is limited by an important domain shift. On the other hand, collecting domain-specific ground truth such as 2D or 3D labels does not scale well. Therefore, we propose a pre-training approach based on self-supervised learning that works on human-centric data using only images. Our method uses pairs of images of humans: the first is partially masked and the model is trained to reconstruct the masked parts given the visible ones and a second image. It relies on both stereoscopic (cross-view) pairs, and temporal (cross-pose) pairs taken from videos, in order to learn priors about 3D as well as human motion. We pre-train a model for body-centric tasks and one for hand-centric tasks. With a generic transformer architecture, these models outperform existing self-supervised pre-training methods on a wide set of human-centric downstream tasks, and obtain state-of-the-art performance for instance when fine-tuning for model-based and model-free human mesh recovery.

We propose Lodge, a network capable of generating extremely long dance sequences conditioned on given music. We design Lodge as a two-stage coarse to fine diffusion architecture, and propose the characteristic dance primitives that possess significant expressiveness as intermediate representations between two diffusion models. The first stage Global diffusion focuses on comprehending the coarse-level music-dance correlation and production characteristic dance primitives. In contrast, the second-stage local diffusion parallelly generates detailed motion sequences under the guidance of the dance primitives and choreographic rules.In addition, we propose a Foot Refine Block to optimize the contact between the feet and the ground, enhancing the physical realism of the motion.Our approach can parallelly generate dance sequences of extremely long length, striking a balance between global choreographic patterns and local motion quality and expressiveness. Extensive experiments validate the efficacy of our algorithm.

We present $\textbf{GALA}$, a framework that takes as input a single-layer clothed 3D human mesh and decomposes it into complete multi-layered 3D assets. The outputs can then be combined with other assets to create novel clothed human avatars with any poses. Existing reconstruction approaches often treat clothed humans as a single-layer of geometry and overlook the inherent compositionality of humans with hairstyles, clothing and accessories, thereby limiting the utility of the meshes for down-stream applications. Decomposing a single-layer mesh into separate layers is a challenging task because it requires the synthesis of plausible geometry and texture for the severely occluded regions. Moreover, even with successful decomposition, meshes are not normalized in terms of poses and body shapes, failing coherent composition with novel identities and poses. To address these challenges, we propose to leverage the general knowledge of a pretrained 2D diffusion model as geometry and appearance prior for humans and other assets. We first separate the input mesh using the 3D surface segmentation extracted from multi-view 2D segmentations. Then we synthesize the missing geometry of different layers in both posed and canonical spaces using a novel pose-guided Score Distillation Sampling (SDS) loss. Once we complete inpainting high-fidelity 3D geometry, we also apply the same SDS loss to its texture to obtain the complete appearance including the initially occluded regions. Through a series of decomposition steps, we obtain multiple layers of 3D assets in a shared canonical space normalized in terms of poses and human shapes, hence effortlessly supporting composition to novel identities and reanimation with novel poses. Our experiments demonstrate the effectiveness of our approach for decomposition, canonicalization, and composition tasks compared to existing solutions.

Recent advances in text-to-motion generation using diffusion and autoregressive models have shown promising results. However, these models often suffer from a trade-off between real-time performance, high fidelity, and motion editability. To address this gap, we introduce MMM, a novel yet simple motion generation paradigm based on Masked Motion Model. MMM consists of two key components: (1) a motion tokenizer that transforms 3D human motion into a sequence of discrete tokens in latent space, and (2) a conditional masked motion transformer that learns to predict randomly masked motion tokens, conditioned on the pre-computed text tokens. By attending to motion and text tokens in all directions, MMM explicitly captures inherent dependency among motion tokens and semantic mapping between motion and text tokens. During inference, this allows parallel and iterative decoding of multiple motion tokens that are highly consistent with fine-grained text descriptions, therefore simultaneously achieving high-fidelity and high-speed motion generation. In addition, MMM has innate motion editability. By simply placing mask tokens in the place that needs editing, MMM automatically fills the gaps while guaranteeing smooth transitions between editing and non-editing parts. Extensive experiments on the HumanML3D and KIT-ML datasets demonstrate that MMM surpasses current leading methods in generating high-quality motion (evidenced by superior FID scores of 0.08 and 0.429), while offering advanced editing features such as body-part modification, motion in-betweening, and the synthesis of long motion sequences. In addition, MMM is two orders of magnitude faster on a single mid-range GPU than editable motion diffusion models. Our project page is available at https://exitudio.github.io/MMM-page/.

Driver's eye gaze holds a wealth of cognitive and intentional cues crucial for intelligent vehicles. Despite its significance, research on in-vehicle gaze estimation remains limited due to the scarcity of comprehensive and well-annotated datasets in real driving scenarios. In this paper, we present three novel elements to advance in-vehicle gaze research. Firstly, we introduce IVGaze, a pioneering dataset capturing in-vehicle gaze, compiled from 125 individuals and covering a large range of gaze and head within vehicles. Conventional gaze collection systems are inadequate for in-vehicle use. In this dataset, we propose a new vision-based solution for in-vehicle gaze collection, introducing a refined gaze target calibration method to tackle annotation challenges. Second, our research focuses on in-vehicle gaze estimation leveraging the IVGaze. Images of in-vehicle faces often suffer from low resolution, prompting our introduction of a gaze pyramid transformer that harnesses transformer-based multilevel features integration. Expanding upon this, we introduce the dual-stream gaze pyramid transformer (GazeDPTR). Employing perspective transformation, we rotate virtual cameras to normalize images, utilizing camera pose to merge normalized and original images for accurate gaze estimation. GazeDPTR showcases state-of-the-art performance on the IVGaze dataset. Thirdly, we explore a novel strategy for gaze zone classification by extending the GazeDPTR. A foundational tri-plane and project gaze onto these planes are newly defined. Leveraging both positional features from the projection points and visual attributes from images, we achieve superior performance compared to relying solely on visual features, thereby substantiating the advantage of gaze estimation. To foster advancements in this domain, the code and dataset will be released to facilitate future research.

This paper addresses the problem of generating lifelike holistic co-speech motions for 3D avatars, focusing on two key aspects: variability and coordination. Variability allows the avatar to exhibit a wide range of motions even with similar speech content, while coordination ensures a harmonious alignment among facial expressions, hand gestures, and body poses. We aim to achieve both with ProbTalk, a unified probabilistic framework designed to jointly model facial, hand, and body movements in speech. ProbTalk builds on the variational autoencoder (VAE) architecture and incorporates three core designs. First, we introduce product quantization (PQ) to the VAE, which enriches the representation of complex holistic motion. Second, we devise a novel non-autoregressive model that embeds 2D positional encoding into the product-quantized representation, thereby preserving essential structure information of the PQ codes. Last, we employ a secondary stage to refine the preliminary prediction, further sharpening the high-frequency details. Coupling these three designs enables ProbTalk to generate natural and diverse holistic co-speech motions, outperforming several state-of-the-art methods in qualitative and quantitative evaluations, particularly in terms of realism. Our code and model will be released for research purposes at https://feifeifeiliu.github.io/probtalk/.

This paper introduces the first text-guided work for generating the sequence of hand-object interaction in 3D. The main challenge arises from the lack of labeled data where existing ground-truth datasets are nowhere near generalizable in interaction type and object category, which inhibits the modeling of diverse 3D hand-object interaction with the correct physical implication (e.g., contacts and semantics) from text prompts. To address this challenge, we propose to decompose the interaction generation task into two subtasks: hand-object contact generation; and hand-object motion generation. For contact generation, a VAE-based network takes as input a text and an object mesh, and generates the probability of contacts between the surfaces of hands and the object during the interaction. The network learns a variety of local geometry structure of diverse objects that is independent of the objects’ category, and thus, it is applicable to general objects. For motion generation, a Transformer-based diffusion model utilizes this 3D contact map as a strong prior for generating physically plausible hand-object motion as a function of text prompts by learning from the augmented labeled dataset; where we annotate text labels from many existing 3D hand and object motion data. Finally, we further introduce a hand refiner module that minimizes the distance between the object surface and hand joints to improve the temporal stability of the object-hand contacts and to suppress the penetration artifacts. In the experiments, we demonstrate that our method can generate more realistic and diverse interactions compared to other baseline methods. We also show that our method is applicable to unseen objects. We will release our model and newly labeled data as a strong foundation for future research. Codes and data are available in: https://github.com/JunukCha/Text2HOI.

While modeling people wearing tight-fitting clothing has made great strides in recent years, loose-fitting clothing remains a challenge. We propose a method that delivers realistic garment models from real-world images, regardless of garment shape or deformation. To this end, we introduce a fitting approach that utilizes shape and deformation priors learned from synthetic data to accurately capture garment shapes and deformations, including large ones. Not only does our approach recover the garment geometry accurately, it also yields models that can be directly used by downstream applications such as animation and simulation.

Information retrieval is an ever-evolving and crucial research domain. The substantial demand for high-quality human motion data especially in online acquirement has led to a surge in human motion research works. Prior works have mainly concentrated on dual-modality learning, such as text and motion tasks, but three-modality learning has been rarely explored. Intuitively, an extra introduced modality can enrich a model’s application scenario, and more importantly, an adequate choice of the extra modality can also act as an intermediary and enhance the alignment between the other two disparate modalities. In this work, we introduce LAVIMO (LAnguage-VIdeo-MOtion alignment), a novel framework for three-modality learning integrating human-centric videos as an additional modality, thereby effectively bridging the gap between text and motion. Moreover, our approach leverages a specially designed attention mechanism to foster enhanced alignment and synergistic effects among text, video, and motion modalities. Empirically, our results on the HumanML3D and KIT-ML datasets show that LAVIMO achieves state-of-the-art performance in various motion-related cross-modal retrieval tasks, including text-to-motion, motion-to-text, video-to-motion and motion-to-video.

We present SplattingAvatar, a hybrid 3D representation of photorealistic human avatars with Gaussian Splatting embedded on a triangle mesh, which renders over 300 FPS on a modern GPU and 30 FPS on a mobile device.We disentangle the motion and appearance of a virtual human with explicit mesh geometry and implicit Gaussian Splatting rendering. The Gaussians are defined by barycentric coordinates and displacement on a triangle mesh as Phong surfaces. We extend lifted optimization to simultaneously optimize the parameters of the Gaussians while walking on the triangle mesh.SplattingAvatar is a hybrid representation of virtual humans where the mesh represents low-frequency motion and surface deformation, while the Gaussians take over the high-frequency geometry and detailed appearance.Unlike existing deformation methods that rely on an MLP-based linear blend skinning (LBS) field for motion, we control the rotation and translation of the Gaussians directly by mesh, which empowers its compatibility with various animation techniques, e.g., skeletal animation, blend shapes, and mesh editing.% Our method can be trained from monocular videos for full-body or head avatars. We demonstrate state-of-the-art quality and real-time rendering on several datasets. We plan to release our source code in the hope of facilitating research on digital humans.Trainable from monocular videos for both full-body and head avatars, SplattingAvatar shows state-of-the-art rendering quality across multiple datasets. We plan to release our source code to support further research in digital human modeling.

Human pose forecasting garners attention for its diverse applications.However, challenges in modeling the multi-modal nature of human motion and intricate interactions among agents persist, particularly with longer timescales and more agents.In this paper, we propose an interaction-aware trajectory-conditioned long-term multi-agent human pose forecasting model, utilizing a coarse-to-fine prediction approach: multi-modal global trajectories are initially forecasted, followed by respective local pose forecasts conditioned on each mode.In doing so, our Trajectory2Pose model introduces a graph-based agent-wise interaction module for a reciprocal forecast of local motion-conditioned global trajectory and trajectory-conditioned local pose.Our model effectively handles the multi-modality of human motion and the complexity of long-term multi-agent interactions, improving performance in complex environments.Furthermore, we address the lack of long-term (6s+) multi-agent (5+) datasets by constructing a new dataset from real-world images and 2D annotations, enabling a comprehensive evaluation of our proposed model.State-of-the-art prediction performance on both complex and simpler datasets confirms the generalized effectiveness of our method.

Recent years have witnessed rapid progress in monocular human mesh recovery.Despite their impressive performance on public benchmarks, existing methods are vulnerable to unusual poses, which prevents them from deploying to challenging scenarios such as dance and martial arts.This issue is mainly attributed to the domain gap induced by the data scarcity in relevant cases.Most existing datasets are captured in constrained scenarios and lack samples of such complex movements. For this reason, we propose a data collection pipeline comprising automatic crawling, precise annotation, and hardcase mining. Based on this pipeline, we establish a large dataset in a short time.The dataset, named HardMo, contains 7M images along with precise annotations covering 15 categories of dance and 14 categories of martial arts.Empirically, we find that the prediction failure in dance and martial arts is mainly characterized by the misalignment of hand-wrist and foot-ankle.To dig deeper into the two hardcases, we leverage the proposed automatic pipeline to filter collected data and construct two subsets named HardMo-Hand and HardMo-Foot. Extensive experiments demonstrate the effectiveness of the annotation pipeline and the data-driven solution to failure cases.Specifically, after being trained on HardMo, HMR, an early pioneering method, can even outperform the current state of the art, 4DHumans, on our benchmarks.

In this work we focus on learning facial representations that can be adapted to train effective face recognition models, particularly in the absence of labels. Firstly, compared with existing labelled face datasets, a vastly larger magnitude of unlabeled faces exists in the real world. We explore the learning strategy of these unlabeled facial images through self-supervised pretraining to transfer generalized face recognition performance. Moreover, motivated by one recent finding, that is, the face saliency area is critical for face recognition, in contrast to utilizing random cropped blocks of images for constructing augmentations in pretraining, we utilize patches localized by extracted facial landmarks. This enables our method - namely Landmark-based Facial Self-supervised learning (LAFS), to learn key representation that is more critical for face recognition. We also incorporate two landmark-specific augmentations which introduce more diversity of landmark information to further regularize the learning. With learned landmark-based facial representations, we further adapt the representation for face recognition with regularization mitigating variations in landmark positions. Our method achieves significant improvement over the state-of-the-art on multiple face recognition benchmarks, especially on more challenging few-shot scenarios.The code is available at https://github.com/szlbiubiubiu/LAFS_CVPR2024

We introduce Motion Diversification Networks, a novel framework for learning to generate realistic and diverse 3D human motion. Despite recent advances in deep generative motion modeling, existing models often fail to produce samples that capture the full range of plausible and natural 3D human motion within a given context. The lack of diversity becomes even more apparent in applications where subtle and multi-modal 3D human forecasting is crucial for safety, such as robotics and autonomous driving. Towards more realistic and functional 3D motion models, we highlight limitations in existing generative modeling techniques, particularly in overly simplistic latent code sampling strategies. We then introduce a transformer-based diversification mechanism that learns to effectively guide sampling in the latent space. Our proposed attention-based module queries multiple stochastic samples to flexibly predict a diverse set of latent codes which can be subsequently decoded into motion samples. The proposed framework achieves state-of-the-art diversity and accuracy prediction performance across a range of benchmarks and settings, particularly when used to forecast intricate in-the-wild 3D human motion within complex urban environments. Our models, datasets, and code are available at https://mdncvpr.github.io/.

Faithfully modeling the space of articulations is a crucial task that allows recovery and generation of realistic poses, and remains a notorious challenge. To this end, we introduce Neural Riemannian Distance Fields (NRDFs), data-driven priors modeling the space of plausible articulations, represented as the zero-level-set of a neural field in a high-dimensional product-quaternion space. To train NRDFs only on positive examples, we introduce a new \textbf{sampling algorithm}, ensuring that the geodesic distances follow a desired distribution, yielding a principled distance field learning paradigm. We then devise a \textbf{projection algorithm} to map any random pose onto the level-set by an \textbf{adaptive-step Riemannian optimizer}, adhering to the product manifold of joint rotations at all times. NRDFs can compute the Riemannian gradient via backpropagation and by mathematical analogy, are related to Riemannian flow matching, a recent generative model. We conduct a comprehensive evaluation of NRDF against other pose priors in various downstream tasks, \emph{i.e.}, pose generation, image-based pose estimation, and solving inverse kinematics, highlighting NRDF's superior performance. Besides humans, NRDF's versatility extends to hand and animal poses, as it can effectively represent any articulation.

3D Morphable Models (3DMMs) provide promising 3D face reconstructions in various applications. However, existing methods struggle to reconstruct faces with extreme expressions due to deficiencies in supervisory signals, such as sparse or inaccurate landmarks. Segmentation information contains effective geometric contexts for face reconstruction. Certain attempts intuitively depend on differentiable renderers to compare the rendered silhouettes of reconstruction with segmentation, which is prone to issues like local optima and gradient instability. In this paper, we fully utilize the facial part segmentation geometry by introducing Part Re-projection Distance Loss (PRDL). Specifically, PRDL transforms facial part segmentation into 2D points and re-projects the reconstruction onto the image plane. Subsequently, by introducing grid anchors and computing different statistical distances from these anchors to the point sets, PRDL establishes geometry descriptors to optimize the distribution of the point sets for face reconstruction. PRDL exhibits a clear gradient compared to the renderer-based methods and presents state-of-the-art reconstruction performance in extensive quantitative and qualitative experiments. Our project is available at https://github.com/wang-zidu/3DDFA-V3.

Most models of visual attention aim at predicting either top-down or bottom-up control, as studied using different visual search and free-viewing tasks. In this paper we propose the Human Attention Transformer (HAT), a single model that predicts both forms of attention control. HAT uses a novel transformer-based architecture and a simplified foveated retina that collectively create a spatio-temporal awareness akin to the dynamic visual working memory of humans. HAT not only establishes a new state-of-the-art in predicting the scanpath of fixations made during target-present and target-absent visual search and "taskless" free viewing, but also makes human gaze behavior interpretable. Unlike previous methods that rely on a coarse grid of fixation cells and experience information loss due to fixation discretization, HAT features a sequential dense prediction architecture and outputs a dense heatmap for each fixation, thus avoiding discretizing fixations. HAT sets a new standard in computational attention, which emphasizes effectiveness, generality, and interpretability. HAT's demonstrated scope and applicability will likely inspire the development of new attention models that can better predict human behavior in various attention-demanding scenarios.

Face Image Quality Assessment (FIQA) is pivotal for guaranteeing the accuracy of face recognition in unconstrained environments. Recent progress in deep quality-fitting-based methods which train models to align with quality anchors, has shown promise in FIQA. However, these methods heavily depend on a recognition model to yield quality anchors and indiscriminately treat the confidence of inaccurate anchors as equivalent to that of accurate ones during the FIQA model training, leading to a fitting bottleneck issue. This paper seeks a solution by putting forward the Confidence-Calibrated Face Image Quality Assessment (CLIB-FIQA) approach, underpinned by the synergistic interplay between the quality anchors and objective quality factors such as blur, pose, expression, occlusion, and illumination. Specifically, we devise a joint learning framework built upon the vision-language alignment model, which leverages the joint distribution with multiple quality factors to facilitate the quality fitting of the FIQA model. Furthermore, to alleviate the issue of the model placing excessive trust in inaccurate quality anchors, we propose a confidence calibration method to correct the quality distribution by exploiting to the fullest extent of these objective quality factors characterized as the merged-factor distribution during training. Experimental results on eight datasets reveal the superior performance of the proposed method. The source code will be made publicly available.

While existing motion style transfer methods are effective between two motions with identical content, their performance significantly diminishes when transferring style between motions with different contents. This challenge lies in the lack of clear separation between content and style of a motion. To tackle this challenge, we propose a novel motion style transformer that effectively disentangles style from content and generates a plausible motion with transferred style from a source motion. Our distinctive approach to achieving the goal of disentanglement is twofold: (1) a new architecture for motion style transformer with 'part-attentive style modulator across body parts' and 'Siamese encoders that encode style and content features separately'; (2) style disentanglement loss. Our method outperforms existing methods and demonstrates exceptionally high quality, particularly in motion pairs with different contents, without the need for heuristic post-processing. Codes are available at https://github.com/Boeun-Kim/MoST.

To adequately utilize the available image evidence in multi-view video-based avatar modeling, we propose TexVocab, a novel avatar representation that constructs a texture vocabulary and associates body poses with texture maps for animation. Given multi-view RGB videos, our method initially back-projects all the available images in the training videos to the posed SMPL surface, producing texture maps in the SMPL UV domain. Then we construct pairs of human poses and texture maps to establish a texture vocabulary for encoding dynamic human appearances under various poses. Unlike the commonly used joint-wise manner, we further design a body-part-wise encoding strategy to learn the structural effects of the kinematic chain. Given a driving pose, we query the pose feature hierarchically by decomposing the pose vector into several body parts and interpolating the texture features for synthesizing fine-grained human dynamics. Overall, our method is able to create animatable human avatars with detailed and dynamic appearances from RGB videos, and the experiments show that our method outperforms state-of-the-art approaches.

In the context of computer vision and human-robot interaction, forecasting 3D human poses is crucial for understanding human behavior and enhancing the predictive capabilities of intelligent systems. While existing methods have made significant progress, they often focus on predicting major body joints, overlooking fine-grained gestures and their interaction with objects. Human hand movements,particularly during object interactions, play a pivotal role and provide more precise expressions of human poses. This work fills this gap and introduces a novel paradigm: forecasting 3D whole-body human poses with a focus on grasping objects. This task involves predicting activities across all joints in the body and hands, encompassing the complexities of internal heterogeneity and external interactivity. To tackle these challenges, we also propose a novel approach: C3HOST, cross-context cross-modal consolidation for 3D whole-body pose forecasting, effectively handles the complexities of internal heterogeneity and external interactivity. C3HOST involves distinct steps, including the heterogeneous content encoding and alignment, and cross-modal feature learning and interaction. These enable us to predict activities across all body and hand joints, ensuring highprecision whole-body human pose prediction, even during object grasping. Extensive experiments on two benchmarks demonstrate that our model significantly enhances the accuracy of whole-body human motion prediction. More results see the supplementary materials.

The advancing of human-scene interaction modeling confronts substantial challenges in the scarcity of high-quality data and advanced motion synthesis methods. Previous endeavors have been inadequate in offering sophisticated datasets that effectively tackle the dual challenges of scalability and data quality. In this work, we overcome these challenges by introducing TRUMANS, a large-scale \mocap dataset created by efficiently and precisely replicating the synthetic scenes in the physical environment. TRUMANS, featuring the most extensive motion-capturedhuman-scene interaction datasets thus far, comprises over 15 hours of diverse human behaviors, including concurrent interactions with dynamic and articulated objects, across 100 indoor scene configurations. It provides accurate pose sequences of both humans and objects, ensuring a high level of contact plausibility during the interaction. Additionally, we also propose a data augmentation approach that automatically adapts collision-free and interaction-precise human motions, significantly increasing the diversity of both interacting objects and scene backgrounds.Leveraging the benefits of TRUMANS, we propose a novel approach that employs a diffusion-based autoregressive mechanism for the real-time generation of human-scene interaction sequences with arbitrary length.The efficacy of TRUMANS and our motion synthesis method is validated through extensive experimental results, surpassing all existing baselines in terms of quality and diversity. Notably, our method demonstrates superb zero-shot generalizability on existing 3D scene datasets eg, PROX, Replica, ScanNet, ScanNet++, capable of generating even more realistic motions than the ground-truth annotations on PROX. Our human study further indicates that our generated motions are almost indistinguishable from the original motion-captured sequences, highlighting their superior quality. Our dataset and model will be released for research purposes.

In recent years, there has been a significant shift in the field of digital avatar research, towards modeling, animating and reconstructing clothed human representations, as a key step towards creating realistic avatars. However, current 3D cloth generation methods are garment specific or trained completely on synthetic data, hence lacking fine details and realism. In this work, we make a step towards automatic realistic garment design and propose Design2Cloth, a high fidelity 3D generative model trained on a real world dataset from more than 2000 subject scans. To provide vital contribution to the fashion industry, we developed a user-friendly adversarial model capable of generating diverse and detailed clothes simply by drawing a 2D cloth mask. Under a series of both qualitative and quantitative experiments, we showcase that Design2Cloth outperforms current state-of-the-art cloth generative models by a large margin. In addition to the generative properties of our network, we showcase that the proposed method can be used to achieve high quality reconstructions from single in-the-wild images and 3D scans. Dataset, code and pre-trained model will become publicly available.

Humans constantly interact with their surrounding environments. Current human-centric generative models mainly focus on synthesizing humans plausibly interacting with $\textit{static}$ scenes and objects, while the $\textit{dynamic}$ human action-reaction synthesis for ubiquitous causal human-human interactions is less explored. Human-human interactions can be regarded as asymmetric with actors and reactors in atomic interaction periods. In this paper, we comprehensively analyze the asymmetric, dynamic, synchronous, and detailed nature of human-human interactions and propose the $\textit{first}$ multi-setting human action-reaction synthesis benchmark to generate human reactions conditioned on given human actions. To begin with, we propose to annotate the actor-reactor order of the interaction sequences for the NTU120, InterHuman and Chi3D datasets. Based on them, a diffusion-based generative model with a Transformer decoder architecture called $\textbf{ReGenNet}$ together with an explicit distance-based interaction loss is proposed to predict human reactions in an online manner, where the future states of actors are unavailable to reactors. Quantitative and qualitative results show that our method can generate instant and plausible human reactions compared to the baselines, and can generalize to unseen actor motions and viewpoint changes.

Reconstructing an avatar from a portrait image has many applications in multimedia, but remains a challenging research problem. Extracting reflectance maps and geometry from one image is ill-posed: recovering geometry is a one-to-many mapping problem, and reflectance and light are difficult to disentangle. Accurate geometry and reflectance can be captured under the controlled conditions of a light stage, but it is costly to acquire large datasets in this fashion. Moreover, training solely with this type of data leads to poor generalization with in-the-wild images. This motivates the introduction of MoSAR, a method for 3D avatar generation from monocular images. We propose a semi-supervised training scheme that improves generalization by learning from both light stage and in-the-wild datasets. This is achieved using a novel differentiable shading formulation. We show that our approach effectively disentangles the intrinsic face parameters, producing relightable avatars. As a result, MoSAR estimates a richer set of skin reflectance maps and generates more realistic avatars than existing state-of-the-art methods. We also release a new dataset, that provides intrinsic face attributes (diffuse, specular, ambient occlusion and translucency maps) for 10k subjects. Project, code and dataset: https://ubisoft-laforge.github.io/character/mosar

3D facial landmark localization has proven to be of particular use for applications, such as face tracking, 3D face modeling and image-based 3D face reconstruction. In the supervised learning case, such methods usually rely on 3D landmark datasets derived from 3DMM-based registration that often lack spatial definition alignment, as compared with that chosen by hand-labeled human consensus, e.g. how are eyebrow landmarks defined? This creates a gap between landmark datasets generated via high-quality 2D human labels and 3DMMs, and it ultimately limits their effectiveness. To address this issue, we introduce a novel semi-supervised learning approach that learns 3D landmarks by directly lifting (visible) hand-labeled 2D landmarks and ensures better definition alignment, without the need for 3D landmark datasets. To lift 2D landmarks to 3D, we leverage 3D-aware GANs for better multi-view consistency learning, and in-the-wild multi-frame videos for robust cross-generalization. Furthermore, we contribute a novel 3D facial landmark evaluation scheme to handle comparison across various 3D landmark definitions by exploiting recent advancements in photogrammetric face mesh tracking. Empirical experiments demonstrate that our method not only achieves better definition alignment between 2D-3D landmarks but also outperforms other supervised learning 3D landmark localization methods on ground-truth 3D datasets.

Domain shift is a challenge for supervised human pose estimation, where the source data and target data come from different distributions. This is why pose estimation methods generally perform worse on the test set than on the training set. Recently, test-time adaptation has proven to be an effective way to deal with domain shift in human pose estimation. Although the performance on the target domain has been improved, existing methods require a large number of weight updates for convergence, which is time-consuming and brings catastrophic forgetting. To solve these issues, we propose a meta-auxiliary learning method to achieve fast adaptation for domain shift during inference. Specifically, we take human pose estimation as the supervised primary task, and propose body-specific image inpainting as a self-supervised auxiliary task. First, we jointly train the primary and auxiliary tasks to get a pre-trained model on the source domain. Then, meta-training correlates the performance of the two tasks to learn a good weight initialization. Finally, meta-testing adapts the meta-learned model to the target data through self-supervised learning. Benefiting from the meta-learning paradigm, the proposed method enables fast adaptation to the target domain while preserving the source domain knowledge. The well-designed auxiliary task better pays attention to human-related semantics in a single image. Extensive experiments demonstrate the effectiveness of our test-time fast adaptation.

We propose FlashAvatar, a novel and lightweight 3D animatable avatar representation that could reconstruct a digital avatar from a short monocular video sequence in minutes and render high-fidelity photo-realistic images at 300FPS on a consumer-grade GPU. To achieve this, we maintain a uniform 3D Gaussian field embedded in the surface of a parametric face model and learn extra spatial offset to model non-surface regions and subtle facial details. While full use of geometric priors can capture high-frequency facial details and preserve exaggerated expressions, proper initialization can help reduce the number of Gaussians, thus enabling super-fast rendering speed. Extensive experimental results demonstrate that FlashAvatar outperforms existing works regarding visual quality and personalized details and is almost an order of magnitude faster in rendering speed. Project page: \href{https://ustc3dv.github.io/FlashAvatar/}{https://ustc3dv.github.io/FlashAvatar/}

Interactive motion synthesis is essential in creating immersive experiences in entertainment applications, such as video games and virtual reality. However, generating animations that are both high-quality and contextually responsive remains a challenge. Traditional techniques in the game industry can produce high-fidelity animations but suffer from high computational costs and poor scalability. Trained neural network models alleviate the memory and speed issues, yet fall short on generating diverse motions. Diffusion models offer diverse motion synthesis with low memory usage, but require expensive reverse diffusion processes. This paper introduces the Accelerated Auto-regressive Motion Diffusion Model (AAMDM), a novel motion synthesis framework designed to achieve quality, diversity, and efficiency all together. AAMDM integrates Denoising Diffusion GANs as a fast Generation Module, and an Auto-regressive Diffusion Model as a Polishing Module. Furthermore, AAMDM operates in a lower-dimensional embedded space rather than the full-dimensional pose space, which reduces the training complexity as well as further improves the performance. We show that AAMDM outperforms existing methods in motion quality, diversity, and runtime efficiency, through comprehensive quantitative analyses and visual comparisons. We also demonstrate the effectiveness of each algorithmic component through ablation studies.

Predicting human pose sequences via existing pose estimators often encounters various estimation errors. Motion refinement methods aim to optimize the predicted human pose sequences from pose estimators while ensuring minimal computational overhead and latency. Prior investigations have primarily concentrated on striking a balance between the two objectives, i.e., smoothness and precision, while optimizing the predicted pose sequences. However, it has come to our attention that the tension between these two objectives can provide additional quality cues about the predicted pose sequences. These cues, in turn, are able to aid the network in optimizing lower-quality poses. To leverage this quality information, we propose a motion refinement network, termed SynSP, to achieve a Synergy of Smoothness and Precision in the sequence refinement tasks. Moreover, SynSP can also address multi-view poses of one person simultaneously, fixing inaccuracies in predicted poses through heightened attention to similar poses from other views, thereby amplifying the resultant quality cues and overall performance. Compared with previous methods, SynSP benefits from both pose quality and multi-view information with a much shorter input sequence length, achieving state-of-the-art results among four challenging datasets involving 2D, 3D, and SMPL pose representations in both single-view and multi-view scenes. We will release our source codes, pretrained models, and online demos to facilitate further research.

Expressive human pose and shape estimation (a.k.a. 3D whole-body mesh recovery) involves the human body, hand, and expression estimation. Most existing methods have tackled this task in a two-stage manner, first detecting the human body part with an off-the-shelf detection model and then inferring the different human body parts individually. Despite the impressive results achieved, these methods suffer from 1) loss of valuable contextual information via cropping, 2) introducing distractions, and 3) lacking inter-association among different persons and body parts, inevitably causing performance degradation, especially for crowded scenes. To address these issues, we introduce a novel all-in-one-stage framework, AiOS, for multiple expressive human pose and shape recovery without an additional human detection step. Specifically, our method is built upon DETR, which treats multi-person whole-body mesh recovery task as a progressive set prediction problem with various sequential detection. We devise the decoder tokens and extend them to our task. Specifically, we first employ a human token to probe a human location in the image and encode global features for each instance, which provides a coarse location for the later transformer block. Then, we introduce a joint-related token to probe the human joint in the image and encoder a fine-grained local feature, which collaborates with the global feature to regress the whole-body mesh. This straightforward but effective model outperforms previous state-of-the-art methods by a 9% reduction in NMVE on AGORA, a 30% reduction in PVE on EHF, a 10% reduction in PVE on ARCTIC, and a 3% reduction in PVE on EgoBody.

Generating a 3D human model from a single reference image is challenging because it requires inferring textures and geometries in invisible views while maintaining consistency with the reference image. Previous methods utilizing 3D generative models are limited by the availability of 3D training data. Optimization-based methods that lift text-to-image diffusion models to 3D generation often fail to preserve the texture details of the reference image, resulting in inconsistent appearances in different views. In this paper, we propose HumanRef, a 3D human generation framework from a single-view input. To ensure the generated 3D model is photorealistic and consistent with the input image, HumanRef introduces a novel method called reference-guided score distillation sampling (Ref-SDS), which effectively incorporates image guidance into the generation process. Furthermore, we introduce region-aware attention to Ref-SDS, ensuring accurate correspondence between different body regions. Experimental results demonstrate that HumanRef outperforms state-of-the-art methods in generating 3D clothed humans with fine geometry, photorealistic textures, and view-consistent appearances. We will make our code and model available upon acceptance.

Generating human motions from textual descriptions has gained growing research interest due to its wide range of applications. However, only a few works consider human-scene interactions together with text conditions, which is crucial for visual and physical realism. This paper focuses on the task of generating human motions in 3D indoor scenes given text descriptions of the human-scene interactions. This task presents challenges due to the multimodality nature of text, scene, and motion, as well as the need for spatial reasoning. To address these challenges, we propose a new approach that decomposes the complex problem into two more manageable sub-problems: (1) language grounding of the target object and (2) object-centric motion generation. For language grounding of the target object, we leverage the power of large language models. For motion generation, we design an object-centric scene representation for the generative model to focus on the target object, thereby reducing the scene complexity and facilitating the modeling of the relationship between human motions and the object. Experiments demonstrate the better motion quality of our approach compared to baselines and validate our design choices. Code will be available at https://zju3dv.github.io/textscenemotion.

Estimating 3D full-body avatars from AR/VR devices is essential for creating immersive experiences in AR/VR applications. This task is challenging due to the limited input from Head Mounted Devices, which capture only sparse observations from the head and hands. Predicting the full-body avatars, particularly the lower body, from these sparse observations presents significant difficulties. In this paper, we are inspired by the inherent property of the kinematic tree defined in the Skinned Multi-Person Linear (SMPL) model, where the upper body and lower body share only one common ancestor node, bringing the potential of decoupled reconstruction. We propose a stratified approach to decouple the conventional full-body avatar reconstruction pipeline into two stages, with the reconstruction of the upper body first and a subsequent reconstruction of the lower body conditioned on the previous stage. To implement this straightforward idea, we leverage the latent diffusion model as a powerful probabilistic generator, and train it to follow the latent distribution of decoupled motions explored by a VQ-VAE encoder-decoder model. Extensive experiments on AMASS mocap dataset demonstrate our state-of-the-art performance in the reconstruction of full-body motions.

We present the Locally Adaptive Morphable Model (LAMM), a highly flexible Auto-Encoder (AE) framework for learning to generate and manipulate 3D meshes. We train our architecture following a simple self-supervised training scheme in which input displacements over a set of sparse control vertices are used to overwrite the encoded geometry in order to transform one training sample into another. During inference, our model produces a dense output that adheres locally to the specified sparse geometry while maintaining the overall appearance of the encoded object. This approach results in state-of-the-art performance in both disentangling manipulated geometry and 3D mesh reconstruction. To the best of our knowledge LAMM is the first end-to-end framework that enables direct local control of 3D vertex geometry in a single forward pass. A very efficient computational graph allows our network to train with only a fraction of the memory required by previous methods and run faster during inference, generating 12k vertex meshes at $>$60fps on a single CPU thread. We further leverage local geometry control as a primitive for higher level editing operations and present a set of derivative capabilities such as swapping and sampling object parts. Code and pretrained models can be found at https://github.com/michaeltrs/LAMM.

We present IntrinsicAvatar, a novel approach to recovering the intrinsic properties of clothed human avatars including geometry, albedo, material, and environment lighting from only monocular videos. Recent advancements in human-based neural rendering have enabled high-quality geometry and appearance reconstruction of clothed humans from just monocular videos. However, these methods bake intrinsic properties such as albedo, material, and environment lighting into a single entangled neural representation. On the other hand, only a handful of works tackle the problem of estimating geometry and disentangled appearance properties of clothed humans from monocular videos. They usually achieve limited quality and disentanglement due to approximations of secondary shading effects via learned MLPs. In this work, we propose to model secondary shading effects explicitly via Monte-Carlo ray tracing. We model the rendering process of clothed humans as a volumetric scattering process, and combine ray tracing with body articulation, essentially tracing secondary rays through the canonical space that represents geometry and appearance. Our approach can recover high-quality geometry, albedo, material, and lighting properties of clothed humans from a single monocular video, without requiring supervised pre-training using ground truth materials which are hard to obtain in practice. Furthermore, since we explicitly model the volumetric scattering process and ray tracing, our model naturally generalizes to novel poses, enabling animation of the reconstructed avatar in novel lighting conditions.

This paper introduces a novel human pose estimation approach using sparse inertial sensors, addressing the shortcomings of previous methods reliant on synthetic data. It leverages a diverse array of real inertial motion capture data from different skeleton formats to improve motion diversity and model generalization. This method features two innovative components: a pseudo-velocity regression model for dynamic motion capture with inertial sensors, and a part-based model dividing the body and sensor data into three regions, each focusing on their unique characteristics. The approach demonstrates superior performance over state-of-the-art models across five public datasets, notably reducing pose error by 19% on the DIP-IMU dataset, thus representing a significant improvement in inertial sensor-based human pose estimation. Our codes are available at https://github.com/dx118/dynaip

We introduce MoMask, a novel masked modeling framework for text-driven 3D human motion generation. In MoMask, a hierarchical quantization scheme is employed to represent human motion as multi-layer discrete motion tokens with high-fidelity details. Starting at the base layer, with a sequence of motion tokens obtained by vector quantization, the residual tokens of increasing orders are derived and stored at the subsequent layers of the hierarchy. This is consequently followed by two distinct bidirectional transformers. For the base-layer motion tokens, a Masked Transformer is designated to predict randomly masked motion tokens conditioned on text input at training stage. During generation (i.e. inference) stage, starting from an empty sequence, our Masked Transformer iteratively fills up the missing tokens; Subsequently, a Residual Transformer learns to progressively predict the next-layer tokens based on the results from current layer. Extensive experiments demonstrate that MoMask outperforms the state-of-art methods on the text-to-motion generation task, with an FID of 0.045 (vs e.g. 0.141 of T2M-GPT) on the HumanML3D dataset, and 0.228 (vs 0.514) on KIT-ML, respectively. MoMask can also be seamlessly applied in related tasks without further model fine-tuning, such as text-guided temporal inpainting.

We propose G-HOP, a denoising diffusion based generative prior for hand-object interactions that allows modeling both the 3D object and a human hand, conditioned on the object category. To learn a 3D spatial diffusion model that can capture this joint distribution, we represent the the human hand via a skeletal distance field to obtain a representation aligned with the (latent) signed distance field for the object. We show that this hand-object prior can then serve as a generic guidance to facilitate other tasks like reconstruction from interaction clip and human grasp synthesis. We believe that our model, trained by aggregating seven diverse real-world interaction datasets spanning across 157 categories, represents a first approach that allows jointly generating both hand and object. Our empirical evaluations demonstrate the benefit from this joint prior in video-based reconstruction and human grasp synthesis, outperforming current task-specific baselines.

Category-agnostic pose estimation (CAPE) aims to predict the pose of a query image based on few support images with pose annotations. Existing methods achieve the localization of arbitrary keypoints through similarity matching between support keypoint features and query image features. However, these methods primarily focus on mining information from the query images, neglecting the fact that support samples with keypoint annotations contain rich category-specific fine-grained semantic information and prior structural information. In this paper, we propose a Support-based Dynamic Perception Network (SDPNet) for the robust and accurate CAPE. On the one hand, SDPNet models complex dependencies between support keypoints, constructing category-specific prior structure to guide the interaction of query keypoints. On the other hand, SDPNet extracts fine-grained semantic information from support samples, dynamically modulating the refinement process of query features. Our method outperforms previous state-of-the-art methods on public datasets by a large margin.

Creating high-fidelity 3D head avatars has always been a research hotspot, but there remains a great challenge under lightweight sparse view setups. In this paper, we propose Gaussian Head Avatar represented by controllable 3D Gaussians for high-fidelity head avatar modeling. We optimize the neutral 3D Gaussians and a fully learned MLP-based deformation field to capture complex expressions. The two parts benefit each other, thereby our method can model fine-grained dynamic details while ensuring expression accuracy. Furthermore, we devise a well-designed geometry-guided initialization strategy based on implicit SDF and Deep Marching Tetrahedra for the stability and convergence of the training procedure. Experiments show our approach outperforms other state-of-the-art sparse-view methods, achieving ultra high-fidelity rendering quality at 2K resolution even under exaggerated expressions.

Existing methods for synthesizing 3D human gestures from speech have shown promising results, but they do not explicitly model the impact of emotions on the generated gestures. Instead, these methods directly output animations from speech without control over the expressed emotion. To address this shortcoming, we present AMUSE, an emotional speech-driven body animation model based on latent diffusion. Our observation is that content (i.e., gestures related to speech rhythm and word utterances), emotion, and personal style are separable. To account for this, AMUSE maps the driving audio to three disentangled latent vectors, one for content, one for emotion and one for personal style. A latent diffusion model, trained to generate gesture motion sequences, is then conditioned on these latent vectors. Once trained, AMUSE synthesizes 3D human gestures directly from speech with control over the emotions and style by combining the content from the driving speech, with the emotion and style of another speech sequence. Randomly sampling the noise of the diffusion model further generates variations of the gesture with the same emotion. Qualitative, quantitative, and perceptual evaluations demonstrate that AMUSE outputs realistic gesture sequences. Compared to the state-of-the-art, the generated gestures are better synchronized with the speech content, and better represent the emotion expressed by the input speech. Code and model will be released for research purposes.

Learning-based approaches to monocular motion capture have recently shown promising results by learning to regress in a data-driven manner. However, due to the challenges in data collection and network designs, it remains challenging for existing solutions to achieve real-time full-body capture while being accurate in world space. In this work, we introduce ProxyCap, a human-centric proxy-to-motion learning scheme to learn world-space motions from a proxy dataset of 2D skeleton sequences and 3D rotational motions. Such proxy data enables us to build a learning-based network with accurate world-space supervision while also mitigating the generalization issues. For more accurate and physically plausible predictions in world space, our network is designed to learn human motions from a human-centric perspective, which enables the understanding of the same motion captured with different camera trajectories. Moreover, a contact-aware neural motion descent module is proposed in our network so that it can be aware of foot-ground contact and motion misalignment with the proxy observations. With the proposed learning-based solution, we demonstrate the first real-time monocular full-body capture system with plausible foot-ground contact in world space even using hand-held moving cameras.

We introduce Multi-view Ancestral Sampling (MAS), a method for 3D motion generation, using 2D diffusion models that were trained on motions obtained from in-the-wild videos. As such, MAS opens opportunities to exciting and diverse fields of motion previously under-explored as 3D data is scarce and hard to collect. MAS works by simultaneously denoising multiple 2D motion sequences representing different views of the same 3D motion. It ensures consistency across all views at each diffusion step by combining the individual generations into a unified 3D sequence, and projecting it back to the original views. We demonstrate MAS on 2D pose data acquired from videos depicting professional basketball maneuvers, rhythmic gymnastic performances featuring a ball apparatus, and horse races. In each of these domains, 3D motion capture is arduous, and yet, MAS generates diverse and realistic 3D sequences. Unlike the Score Distillation approach, which optimizes each sample by repeatedly applying small fixes, our method uses a sampling process that was constructed for the diffusion framework. As we demonstrate, MAS avoids common issues such as out-of-domain sampling and mode-collapse.

3D head avatars built with neural implicit volumetric representations have achieved unprecedented levels of photorealism. However, the computational cost of these methods remains a significant barrier to their widespread adoption, particularly in real-time applications such as virtual reality and teleconferencing. While attempts have been made to develop fast neural rendering approaches for static scenes, these methods cannot be simply employed to support realistic facial expressions, such as in the case of a dynamic facial performance. To address these challenges, we propose a novel fast 3D neural implicit head avatar model that achieves real-time rendering while maintaining fine-grained controllability and high rendering quality. Our key idea lies in the introduction of local hash table blendshapes, which are learned and attached to the vertices of an underlying face parametric model. These per-vertex hash-tables are linearly merged with weights predicted via a CNN, resulting in expression dependent embeddings. Our novel representation enables efficient density and color predictions using a lightweight MLP, which is further accelerated by a hierarchical nearest neighbor search method. Extensive experiments show that our approach runs in real-time while achieving comparable rendering quality to state-of-the-arts and decent results on challenging expressions.

Sign Languages (SL) serve as the predominant mode of communication for the Deaf and Hard of Hearing communities. The advent of deep learning has aided numerous methods in SL recognition and translation, achieving remarkable results. However, Sign Language Production (SLP) poses a challenge for the computer vision community as the motions generated must be realistic and have precise semantic meanings. Most SLP methods rely on 2D data, thus impeding their ability to attain a necessary level of realism. In this work, we propose a diffusion-based SLP model trained on a curated large-scale dataset of 4D signing avatars and their corresponding text transcripts. The proposed method can generate dynamic sequences of 3D avatars from an unconstrained domain of discourse using a diffusion process formed on a novel and anatomically informed graph neural network defined on the SMPL-X body skeleton. Through a series of quantitative and qualitative experiments, we show that the proposed method considerably outperforms previous methods of SLP. We believe that this work presents an important and necessary step towards realistic neural sign avatars, bridging the communication gap between Deaf and hearing communities. The code, method and generated data will be made publicly available.

This paper focuses on advancing the applicability of human avatar learning methods by proposing RAM-Avatar, which learns a Real-time, photo-realistic Avatar that supports full-body control from Monocular videos. To achieve this goal, RAM-Avatar leverages two statistical templates responsible for modeling the facial expression and hand gesture variations, while a sparsely computed dual attention module is introduced upon another body template to facilitate high-fidelity texture rendering for the torsos and limbs. Building on this foundation, we deploy a lightweight yet powerful StyleUnet along with a temporal-aware discriminator to achieve real-time realistic rendering. To enable robust animation for out-of-distribution poses, we propose a Motion Distribution Align module to compensate for the discrepancies between the training and testing motion distribution. Results and extensive experiments conducted in various experimental settings demonstrate the superiority of our proposed method, and a real-time live system is proposed to further push research into applications. The training and testing code will be released for research purposes.

Gaze is a powerful form of non-verbal communication that humans develop from an early age. As such, modeling this behavior is an important task that can benefit a broad set of application domains ranging from robotics to sociology. In particular, the gaze following task in computer vision is defined as the prediction of the 2D pixel coordinates where a person in the image is looking. Previous attempts in this area have primarily centered on CNN-based architectures, but they have been constrained by the need to process one person at a time, which proves to be highly inefficient. In this paper, we introduce a novel and effective multi-person transformer-based architecture for gaze prediction. While there exist prior works using transformers for multi-person gaze prediction (Tu et al., Tonini et al.), they use a fixed set of learnable embeddings to decode both the person and its gaze target, which requires a matching step afterward to link the predictions with the annotations. Thus, it is difficult to quantitatively evaluate these methods reliably with the available benchmarks, or integrate them into a larger human behavior understanding system. Instead, we are the first to propose a multi-person transformer-based architecture that maintains the original task formulation and ensures control over the people fed as input. Our main contribution lies in encoding the person-specific information into a single controlled token to be processed alongside image tokens and using its output for prediction based on a novel multiscale decoding mechanism. Our new architecture achieves state-of-the-art results on the GazeFollow, VideoAttentionTarget, and ChildPlay datasets and significantly outperforms existing multi-person architectures. Our code, checkpoints, and other artifacts will be made publicly available upon acceptance.

Understanding the dynamics of generic 3D scenes is fundamentally challenging in computer vision, essential in enhancing applications related to scene reconstruction, motion tracking, and avatar creation. In this work, we address the task as the problem of inferring dense, long-range motion of 3D points. By observing a set of point trajectories, we aim to learn an implicit motion field parameterized by a neural network to predict the movement of novel points within the same domain, without relying on any data-driven or scene-specific priors. To achieve this, our approach builds upon the recently introduced dynamic point field model that learns smooth deformation fields between the canonical frame and individual observation frames. However, temporal consistency between consecutive frames is neglected, and the number of required parameters increases linearly with the sequence length due to per-frame modeling. To address these shortcomings, we exploit the intrinsic regularization provided by SIREN, and modify the input layer to produce a spatiotemporally smooth motion field. Additionally, we analyze the motion field Jacobian matrix, and discover that the motion degrees of freedom (DOFs) in an infinitesimal area around a point and the network hidden variables have different behaviors to affect the model's representational power. This enables us to improve the model representation capability while retaining the model compactness. Furthermore, to reduce the risk of overfitting, we introduce a regularization term based on the assumption of piece-wise motion smoothness. Our experiments assess the model's performance in predicting unseen point trajectories and its application in temporal mesh alignment with guidance. The results demonstrate its superiority and effectiveness. The code and data for the project are publicly available at https://yz-cnsdqz.github.io/eigenmotion/DOMA.

The authentic 3D hand avatar with every identifiable information, such as hand shapes and textures, is necessary for immersive experiences in AR/VR. In this paper, we present a universal hand model (UHM), which 1) can universally represent high-fidelity 3D hand meshes of arbitrary identities (IDs) and 2) can be adapted to each person with a short phone scan for the authentic hand avatar. For effective universal hand modeling, we perform tracking and modeling at the same time, while previous 3D hand models perform them separately. The conventional separate pipeline suffers from the accumulated errors from the tracking stage, which cannot be recovered in the modeling stage. On the other hand, ours does not suffer from the accumulated errors while having a much more concise overall pipeline. We additionally introduce a novel image matching loss function to address a skin sliding during the tracking and modeling, while existing works have not focused on it much. Finally, using learned priors from our UHM, we effectively adapt our UHM to each person's short phone scan for the authentic hand avatar.

Human image editing includes tasks like changing a person's pose, their clothing, or editing the image according to a text prompt. However, prior work often tackles these tasks separately, overlooking the benefit of mutual reinforcement from learning them jointly. In this paper, we propose UniHuman, a unified model that addresses multiple facets of human image editing in real-world settings. To enhance the model's generation quality and generalization capacity, we leverage guidance from human visual encoders and introduce a lightweight pose-warping module that can exploit different pose representations, accommodating unseen textures and patterns. Furthermore, to bridge the disparity between existing human editing benchmarks with real-world data, we curated 400K high-quality human image-text pairs for training and collected 2K human images for out-of-domain testing, both encompassing diverse clothing styles, backgrounds, and age groups. Experiments on both in-domain and out-of-domain test sets demonstrate that UniHuman outperforms task-specific models by a significant margin. In user studies, UniHuman is preferred by the users in an average of 77\% of cases.

Graph Convolutional Networks (GCNs) have long set the state-of-the-art in skeleton-based action recognition, leveraging their ability to unravel the complex dynamics of human joint topology through the graph’s adjacency matrix. However, an inherent flaw has come to light in these cutting-edge models: they tend to optimize the adjacency matrix jointly with the model weights. This process, while seemingly efficient, causes a gradual decay of bone connectivity data, resulting in a model indifferent to the very topology it sought to represent. To remedy this, we propose a two-fold strategy: (1) We forge an innovative approach that encodes bone connectivity by harnessing the power of graph distances to describe the physical topology; we further incorporate action-specific topological representation via persistent homology analysis to depict systemic dynamics. This preserves the vital topological nuances often lost in conventional GCNs. (2) Our investigation also reveals the redundancy in existing GCNs for multi-relational modeling, which we address by proposing an efficient refinement to Graph Convolutions (GC) - the BlockGC. This significantly reduces parameters while improving performance beyond original GCNs. Our full model, the BlockGCN, establishes new benchmarks in skeleton-based action recognition across all model categories. Its high accuracy and lightweight design, most notably on the large-scale NTU RGB+D 120 dataset, stand as strong validation of the efficacy of BlockGCN.

We introduce GaussianAvatar, a novel approach for real-time, memory-efficient, high-quality animatable human modeling. GaussianAvatar takes as input a single monocular video to create a digital avatar capable of re-articulation in new poses and real-time rendering from novel viewpoints, while seamlessly integrating with rasterization-based graphics pipelines. Central to our method is the Gaussians-on-Mesh representation, a hybrid 3D model combining rendering quality and speed of Gaussian splatting with geometry modeling and compatibility of deformable meshes. We assess GaussianAvatar on ZJU-MoCap data and various YouTube videos. GaussianAvatar matches or surpasses current monocular human modeling algorithms in rendering quality and significantly outperforms them in computational efficiency (43 FPS) while being memory-efficient (3.63 MB per subject).

The estimation of 3D human motion from video has progressed rapidly but current methods still have several key limitations. First, most methods estimate the human in camera coordinates. Second, prior work on estimating humans in global coordinates often assumes a flat ground plane and produces foot sliding. Third, the most accurate methods rely on computationally expensive optimization pipelines, limiting their use to offline applications. Finally, existing video-based methods are surprisingly less accurate than single-frame methods. We address these limitations with WHAM (World-grounded Humans with Accurate Motion), which accurately and efficiently reconstructs 3D human motion in a global coordinate system from video. WHAM learns to lift 2D keypoint sequences to 3D using motion capture data and fuses this with video features, integrating motion context and visual information. WHAM exploits camera angular velocity estimated from a SLAM method together with human motion to estimate the body's global trajectory. We combine this with a contact-aware trajectory refinement method that lets WHAM capture human motion in diverse conditions, such as climbing stairs. WHAM outperforms all existing 3D human motion recovery methods across multiple in-the-wild benchmarks. Code will be available for research purposes.

Self-supervised pre-training has been proved to be effective in learning transferable representations that benefit various visual tasks. This paper asks this question: can self-supervised pre-training learn general facial representations for various facial analysis tasks? Recent efforts toward this goal are limited to treating each face image as a whole, i.e., learning consistent facial representations at the image-level, which overlooks the consistency of local facial representations (i.e., facial regions like eyes, nose, etc). In this work, we make a first attempt to propose a novel self-supervised facial representation learning framework to learn consistent global and local facial representations, Facial Region Awareness (FRA). Specifically, we explicitly enforce the consistency of facial regions by matching the local facial representations across views, which are extracted with learned heatmaps highlighting the facial regions. Inspired by the mask prediction in supervised semantic segmentation, we obtain the heatmaps via cosine similarity between the per-pixel projection of feature maps and facial mask embeddings computed from learnable positional embeddings, which leverage the attention mechanism to globally look up the facial image for facial regions. To learn such heatmaps, we formulate the learning of facial mask embeddings as a deep clustering problem by assigning the pixel features from the feature maps to them. The transfer learning results on facial classification and regression tasks show that our FRA outperforms previous pre-trained models and more importantly, using ResNet as the unified backbone for various tasks, our FRA achieves comparable or even better performance compared with SOTA methods in facial analysis tasks.

We introduce PoseGPT, a framework employing Large Language Models (LLMs) to understand and reason about 3D human poses from images or textual descriptions. Our work is motivated by the human ability to intuitively understand postures from a single image or a brief description, a process that intertwines image interpretation, world knowledge, and an understanding of body language. Traditional human pose estimation methods, whether image-based or text-based, often lack holistic scene comprehension and nuanced reasoning, leading to a disconnect between visual data and its real-world implications. PoseGPT addresses these limitations by embedding SMPL poses as a distinct signal token within a multi-modal LLM, enabling direct generation of 3D body poses from both textual and visual inputs. This approach not only simplifies pose prediction but also empowers LLMs to apply their world knowledge in reasoning about human poses, fostering two advanced tasks: speculative pose generation and reasoning about pose estimation. These tasks involve generating human poses from subtle text queries, possibly accompanied by images, after comprehensive reasoning. We establish benchmarks for these tasks, moving beyond the confines of traditional pose generation and estimation methodologies. Our results show that PoseGPT outperforms existing multimodal LLMs and task-sepcific methods on these newly proposed tasks. Furthermore, PoseGPT's ability to understand and generate 3D human poses based on complex reasoning opens new directions in human pose analysis. We will release the models and training code for research purposes.

Facial action unit (AU) intensity plays a pivotal role in quantifying fine-grained expression behaviors, which is an effective condition for facial expression manipulation. However, publicly available datasets containing intensity annotations for multiple AUs remain severely limited, often featuring a restricted number of subjects. This limitation places challenges to the AU intensity manipulation in images due to disentanglement issues, leading researchers to resort to other large datasets with pretrained AU intensity estimators for pseudo labels.In addressing this constraint and fully leveraging manual annotations of AU intensities for precise manipulation, we introduce AUEditNet.Our proposed model achieves impressive intensity manipulation across 12 AUs, trained effectively with only 18 subjects. Utilizing a dual-branch architecture, our approach achieves comprehensive disentanglement of facial attributes and identity without necessitating additional loss functions or implementing with large batch sizes.This approach offers a potential solution to achieve desired facial attribute editing despite the dataset’s limited subject count.Our experiments demonstrate AUEditNet's superior accuracy in editing AU intensities, affirming its capability in disentangling facial attributes and identity within a limited subject pool. AUEditNet allows conditioning by either intensity values or target images, eliminating the need for constructing AU combinations for specific facial expression synthesis. Moreover, AU intensity estimation, as a downstream task, validates the consistency between real and edited images, confirming the effectiveness of our proposed AU intensity manipulation method.

In human-centric content generation, the pre-trained text-to-image models struggle to produce user-wanted portrait images, which retain the identity of individual while exhibit diverse expressions. This paper introduces our efforts towards the personalized face generation. To this end, we propose a novel multi-modal face generation framework, capable of simultaneous identity-expression control and more fine-grained expression synthesis. Our expression control is so sophisticated that it can be specialized by the fine-grained emotional vocabulary. We devise a novel diffusion model which can undertake the task of simultaneously face swapping and reenactment. Due to the entanglement of identity and expression, it's nontrivial to separately and precisely control them in one framework, thus has not been explored yet. To overcome this, we propose several innovative designs in conditional diffusion model, including balancing identity and expression encoder, improved midpoint sampling and explicitly background conditioning. Extensive experiments have demonstrated the controllability and scalability of the proposed framework, in comparison with state-of-the-art text-to-image, face swapping and face reenactment methods.

Camera-parameter-free multi-view pose estimation is an emerging technique for 3D human pose estimation (HPE). They can infer the camera settings implicitly or explicitly to mitigate the depth uncertainty impact, showcasing significant potential in real applications. However, due to the limited camera setting diversity in the available datasets, the inferred camera parameters are always simply hardcoded into the model during training and not adaptable to the input in inference, making the learned models cannot generalize well under unseen camera settings. A natural solution is to artificially synthesize some samples, i.e., 2D-3D pose pairs, under massive new camera settings. Unfortunately, to prevent over-fitting the existing camera setting, the number of synthesized samples for each new camera setting should be comparable with that for the existing one, which multiplies the scale of training and even makes it computationally prohibitive. In this paper, we propose a novel HPE approach under the invariant risk minimization (IRM) paradigm. Precisely, we first synthesize 2D poses from myriad camera settings. We then train our model under the IRM paradigm, which targets at learning a common optimal model across all camera settings and thus enforces the model to automatically learn the camera parameters based on the input data. This allows the model to accurately infer 3D poses on unseen data by training on only a handful of samples from each synthesized setting and thus avoid the unbearable training cost increment. Another appealing feature of our method is that benefited from the capabilty of IRM in identifying the invariant features, its performance on the seen camera settings is enhanced as well. Comprehensive experiments on both Human3.6M and TotalCapture datasets clearly attest to the superiority of our approach.

Long-term and accurate forecasting is the long-standing pursuit of the human motion prediction task. Existing methods typically suffer from dramatic degradation in prediction accuracy with the increasing prediction horizon. It comes down to two reasons: 1) Insufficient numerical stability. Unforeseen high noise and complex feature relationships in the data. 2) Inadequate modeling stability. Unreasonable step sizes and undesirable parameter updates in the prediction.In this paper, we design a novel and symplectic integral-inspired framework named symplectic integral neural network (SINN), which engages symplectic trajectories to optimize the pose representation and employs a stable symplectic operator to alternately model the dynamic context. Specifically, we design a Symplectic Representation Encoder that performs on enhanced human pose representation to obtain trajectories on the symplectic manifold, ensuring numerical stability based on Hamiltonian mechanics and symplectic spatial splitting algorithm. We further present the Symplectic Temporal Aggregation module in the light of the symplectic temporal splitting algorithm, which splits the long-term prediction into multiple accurate short-term predictions generated by a symplectic operator to secure modeling stability. Moreover, our approach is model-agnostic and can be efficiently integrated with different physical dynamics models.The experimental results demonstrate that our method achieves the new state-of-the-art, outperforming existing methods by large margins: 20.1% on Human3.6M, 16.7% on CUM Mocap, and 10.2% on 3DPW. Our code is anonymously released.

Predicting future human pose is a fundamental application for machine intelligence, which drives robots to plan their behavior and paths ahead of time to seamlessly accomplish human-robot collaboration in real-world 3D scenarios. Despite encouraging results, existing approaches rarely consider the effects of the external scene on the motion sequence, leading to pronounced artifacts and physical implausibilities in the predictions. To address this limitation, this work introduces a novel multi-modal sense-informed motion prediction approach, which conditions high-fidelity generation on two modal information: external 3D scene, and internal human gaze, and is able to recognize their salience for future human activity. Furthermore, the gaze information is regarded as the human intention, and combined with both motion and scene features, we construct a ternary intention-aware attention to supervise the generation to match where the human wants to reach. Meanwhile, we introduce semantic coherence-aware attention to explicitly distinguish the salient point clouds and the underlying ones, to ensure a reasonable interaction of the generated sequence with the 3D scene. On two real-world benchmarks, the proposed method achieves state-of-the-art performance both in 3D human pose and trajectory prediction. More detailed results are available on the page: https://sites.google.com/view/cvpr2024sif3d.

Capturing and preserving motion semantics is essential to motion retargeting between animation characters. However, most of the previous works neglect the semantic information or rely on human-designed joint-level representations. Here, we present a novel Semantics-aware Motion reTargeting (SMT) method with the advantage of vision-language models to extract and maintain meaningful motion semantics. We utilize a differentiable module to render 3D motions. Then the high-level motion semantics are incorporated into the motion retargeting process by feeding the vision-language model with the rendered images and aligning the extracted semantic embeddings. To ensure the preservation of fine-grained motion details and high-level semantics, we adopt a two-stage pipeline consisting of skeleton-aware pre-training and fine-tuning with semantics and geometry constraints. Experimental results show the effectiveness of the proposed method in producing high-quality motion retargeting results while accurately preserving motion semantics. Project page can be found at https://sites.google.com/view/smtnet.

In this work, we introduce two types of makeup prior models to extend existing 3D face prior models: PCA-based and StyleGAN2-based priors. The PCA-based prior model is a linear model that is easy to construct and is computationally efficient. However, it retains only low-frequency information. Conversely, the StyleGAN2-based model can represent high-frequency information with relatively higher computational cost than the PCA-based model. Although there is a trade-off between the two models, both are applicable to 3D facial makeup estimation and related applications. By leveraging makeup prior models and designing a makeup consistency module, we effectively address the challenges that previous methods faced in estimating makeup, particularly in the context of large poses such as oblique faces and subtle light makeup. In experiments, we demonstrate that our approach reduces computational costs by several orders of magnitude, achieving speeds up to 180 times faster. In addition, by improving the accuracy of the estimated makeup, we confirm that our methods are highly advantageous for various 3D facial makeup applications such as 3D makeup face reconstruction, user-friendly makeup editing, makeup transfer, and interpolation.

We propose FaceCom, a method for 3D facial shape completion, which delivers high-fidelity results for incomplete facial inputs of arbitrary forms. Unlike end-to-end shape completion methods based on point clouds or voxels, our approach relies on a mesh-based generative network that is easy to optimize, enabling it to handle shape completion for irregular facial scans. We first train a shape generator on a mixed 3D facial dataset containing 2405 identities. Based on the incomplete facial input, we fit complete faces using an optimization approach under image inpainting guidance. The completion results are refined through a post-processing step. FaceCom demonstrates the ability to effectively and naturally complete facial scan data with varying missing regions and degrees of missing areas. Our method can be used in medical prosthetic fabrication and the registration of deficient scanning data. Our experimental results demonstrate that FaceCom achieves exceptional performance in fitting and shape completion tasks.

Text descriptions intended to guide the facial attributes of the synthesized face may fall short, owing to the intricate entanglement of identity information with identity-irrelevant facial attributes derived from the reference image. To address these issues, we present the novel use of the extended StyleGAN embedding space $\mathcal{W}_+$, to achieve enhanced identity preservation and disentanglement for diffusion models. By aligning this semantically meaningful human face latent space with text-to-image diffusion models, we succeed in maintaining high fidelity in identity preservation, coupled with the capacity for semantic editing. Additionally, we propose new training objectives to balance the influences of both prompt and identity conditions, ensuring that the identity-irrelevant background remains unaffected during facial attribute modifications. Extensive experiments reveal that our method adeptly generates personalized text-to-image outputs that are not only compatible with prompt descriptions but also amenable to common StyleGAN editing directions in diverse settings. Our source code will be made publicly available.

Understanding how we grasp objects with our hands has important applications in areas like robotics and mixed reality.However, this challenging problem requires accurate modeling of the contact between hands and objects.To capture grasps, existing methods use skeletons, meshes, or parametric models that can cause misalignments resulting in inaccurate contacts. We present MANUS, a method for MArkerless Hand-Object Grasp Capture using Articulated 3D GaussiansWe build a novel articulated 3D Gaussians representation that extends 3D Gaussian splatting for high-fidelity representation of articulating hands. Since our representation uses Gaussian primitives, it enables us to efficiently and accurately estimate contacts between the hand and the object. For the most accurate results, our method requires tens of camera views that current datasets do not provide.We therefore build MANUS-Grasps, a new dataset that contains hand-object grasps viewed from 53 cameras across 30+ scenes, 3 subjects, and comprising over 7M frames. In addition to extensive qualitative results, we also show that our method outperforms others on a quantitative contact evaluation method that uses paint transfer from the object to the hand.

Existing wearable motion capture methods typically demand tight on-body fixation (often using straps) for reliable sensing, limiting their application in everyday life.In this paper, we introduce Loose Inertial Poser, a novel motion capture solution with high wearing comfortableness, by integrating four Inertial Measurement Units (IMUs) into a loose-wear jacket. Specifically, we address the challenge of scarce loose-wear IMU training data by proposing a Secondary Motion AutoEncoder (SeMo-AE) that learns to model and synthesize the effects of secondary motion between the skin and loose clothing on IMU data. Our SeMo-AE consists of two novel techniques: i) noise-guided latent space learning, which reverses the process of noise modeling to instead model the latent space, thus enabling easy extrapolation of secondary motion effects; ii) a temporal coherence strategy that models the continuity of secondary motions across successive frames.SeMo-AE is then leveraged to generate a diverse synthetic dataset of loose-wear IMU data to augment training for the pose estimation network and significantly improve its accuracy. For validation, we collected a dataset with various subjects and 2 wearing styles (zipped and unzipped). Experimental results demonstrate that our approach maintains high-quality real-time posture estimation even in loose-wear scenarios.

Coordinate based implicit neural representations have gained rapid popularity in recent years as they have been successfully used in image, geometry and scene modelling tasks. In this work, we present a novel use case for such implicit representations in the context of learning anatomically constrained face models. Actor specific anatomically constrained face models are the state of the art in both facial performance capture and performance retargeting. Despite their practical success, these anatomical models are slow to evaluate and often require extensive data capture to be built. We propose the anatomical implicit face model; an ensemble of implicit neural networks that jointly learn to model the facial anatomy and the skin surface with high fidelity, and can readily be used as a drop in replacement to conventional blendshape models. Given an arbtrary set of skin surface meshes of an actor and only a neutral shape with estimated skull and jaw bones, our method can recover a dense anatomical substructure which constrains every point on the facial surface. We demonstrate the usefulness of our approach in several tasks ranging from shape fitting, shape editing, and performance retargeting.

This paper presents the DiffusionRegPose, a novel approach to multi-person pose estimation that converts a one-stage, end-to-end keypoint regression model into a diffusion-based sampling process. Existing one-stage deterministic regression methods, though efficient, are often prone to missed or false detections in crowded or occluded scenes, due to their inability to reason pose ambiguity. To address these challenges, we handle ambiguous poses in a generative fashion, i.e., sampling from the image-conditioned pose distributions characterized by a diffusion probabilistic model. Specifically, with initial pose tokens extracted from the image, noisy pose candidates are progressively refined by interacting with the initial tokens via attention layers. Extensive evaluations on the COCO and CrowdPose datasets show that DiffusionRegPose clearly improves the pose accuracy in crowded scenarios, as evidenced by a notable 3.3 AP increase in the $AP_H$ metric on the CrowdPose dataset. This demonstrates the model's potential for robust and precise human pose estimation in real-world applications.

3D human pose data collected in controlled laboratory settings present challenges for pose estimators that generalize across diverse scenarios. To address this, domain generalization is employed. Current methodologies in domain generalization for 3D human pose estimation typically utilize adversarial training to generate synthetic poses for training. Nonetheless, these approaches exhibit several limitations. First, the lack of prior information about the target domain complicates the application of suitable augmentation through a single pose augmentor, affecting generalization on target domains. Moreover, adversarial training's discriminator tends to enforce similarity between source and synthesized poses, impeding the exploration of out-of-source distributions. Furthermore, the pose estimator's optimization is not exposed to domain shifts, limiting its overall generalization ability.To address these limitations, we propose a novel framework featuring two pose augmentors: the weak and the strong augmentors. Our framework employs differential strategies for generation and discrimination processes, facilitating the preservation of knowledge related to source poses and the exploration of out-of-source distributions without prior information about target poses. Besides, we leverage meta-optimization to simulate domain shifts in the optimization process of the pose estimator, thereby improving its generalization ability. Our proposed approach significantly outperforms existing methods, as demonstrated through comprehensive experiments on various benchmark datasets.

Comprehensive capturing of human motions requires both accurate captures of complex poses and precise localization of the human within scenes. Most of the HPE datasets and methods primarily rely on RGB, LiDAR, or IMU data. However, solely using these modalities or a combination of them may not be adequate for HPE, particularly for complex and fast movements. For holistic human motion understanding, we present RELI11D, a high-quality multimodal human motion dataset involves LiDAR, IMU system, RGB camera, and Event camera. It records the motions of 10 actors performing 5 sports in 7 scenes, including 3.32 hours of synchronized LiDAR point clouds, IMU measurement data, RGB videos and Event steams. Through extensive experiments, we demonstrate that the RELI11D presents considerable challenges and opportunities as it contains many rapid and complex motions that require precise location. To address the challenge of integrating different modalities, we propose LEIR, a multimodal baseline that effectively utilizes LiDAR Point Cloud, Event stream, and RGB through our cross-attention fusion strategy. We show that LEIR exhibits promising results for rapid motions and daily motions and that utilizing the characteristics of multiple modalities can indeed improve HPE performance. Both the dataset and source code will be released publicly to the research community, fostering collaboration and enabling further exploration in this field.

Co-speech gestures, if presented in the lively form of videos, can achieve superior visual effects in human-machine interaction. While previous works mostly generate structural human skeletons, resulting in the omission of appearance information, we focus on the direct generation of audio-driven co-speech gesture videos in this work. There are two main challenges: 1) A suitable motion feature is needed to describe complex human movements with crucial appearance information. 2) Gestures and speech exhibit inherent dependencies and should be temporally aligned even of arbitrary length. To solve these problems, we present a novel motion-decoupled framework to generate co-speech gesture videos. Specifically, we first introduce a well-designed nonlinear TPS transformation to obtain latent motion features preserving essential appearance information. Then a transformer-based diffusion model is proposed to learn the temporal correlation between gestures and speech, and performs generation in the latent motion space, followed by an optimal motion selection module to produce long-term coherent and consistent gesture videos. For better visual perception, we further design a refinement network focusing on missing details of certain areas. Extensive experimental results show that our proposed framework significantly outperforms existing approaches in both motion and video-related evaluations. Our code, demos, and more resources are available at https://github.com/thuhcsi/S2G-MDDiffusion.

Extracting keypoint locations from input hand frames, known as 3D hand pose estimation, is a critical task in various human-computer interaction applications. Essentially, the 3D hand pose estimation can be regarded as a 3D point subset generative problem conditioned on input frames. Thanks to the recent significant progress on diffusion-based generative models, hand pose estimation can also benefit from the diffusion model to estimate keypoint locations with high quality. However, directly deploying the existing diffusion models to solve hand pose estimation is non-trivial, since they cannot achieve the complex permutation mapping and precise localization. Based on this motivation, this paper proposes HandDiff, a diffusion-based hand pose estimation model that iteratively denoises accurate hand pose conditioned on hand-shaped image-point clouds. In order to recover keypoint permutation and accurate location, we further introduce joint-wise condition and local detail condition. Experimental results show that the proposed model significantly outperforms the existing methods on three hand pose benchmark datasets. Codes and pre-trained models are publicly available at https://anonymous.4open.science/r/HandDiff_-A032.

Normalizing flows have proven their efficacy for density estimation in Euclidean space, but their application to rotational representations, crucial in various domains such as robotics or human pose modeling, remains underexplored. Probabilistic models of the human pose can benefit from approaches that rigorously consider the rotational nature of human joints. For this purpose, we introduce HuProSO3, a normalizing flow model that operates on a high-dimensional product space of SO(3) manifolds, modeling the joint distribution for human joints with three degrees of freedom. HuProSO3's advantage over state-of-the-art approaches is demonstrated through its superior modeling accuracy in three different applications. This work not only addresses the technical challenge of learning densities on SO(3) manifolds, but it also has broader implications for domains where the probabilistic regression of correlated 3D rotations is of importance.

3D pose transfer that aims to transfer the desired pose to a target mesh is one of the most challenging 3D generation tasks. Previous attempts rely on well-defined parametric human models or skeletal joints as driving pose sources. However, to obtain those clean pose sources, cumbersome but necessary pre-processing pipelines are inevitable, hindering implementations of the real-time applications. This work is driven by the intuition that the robustness of the model can be enhanced by introducing adversarial samples into the training, leading to a more invulnerable model to the noisy inputs, which even can be further extended to directly handling the real-world data like raw point clouds/scans without intermediate processing. Furthermore, we propose a novel 3D pose Masked Autoencoder (3D-PoseMAE), a customized MAE that effectively learns 3D extrinsic presentations (i.e., pose). 3D-PoseMAE facilitates learning from the aspect of extrinsic attributes by simultaneously generating adversarial samples that perturb the model and learning the arbitrary raw noisy poses via a multi-scale masking strategy. Both qualitative and quantitative studies show that the transferred meshes given by our network result in much better quality. Besides, we demonstrate the strong generalizability of our method on various poses, different domains, and even raw scans. Experimental results also show meaningful insights that the intermediate adversarial samples generated in the training can successfully attack the existing pose transfer models.

While current methods have shown promising progress on estimating 3D human motion from monocular videos, their motion estimates are often physically unrealistic because they mainly consider kinematics. In this paper, we introduce Physics-aware Pretrained Transformer (PhysPT), which improves kinematics-based motion estimates and infers motion forces. PhysPT exploits a Transformer encoder-decoder backbone to effectively learn human dynamics in a self-supervised manner. Moreover, it incorporates physics principles governing human motion. Specifically, we build a physics-based body representation and contact force model. We leverage them to impose novel physics-inspired training losses (i.e., force loss, contact loss, and Euler-Lagrange loss), enabling PhysPT to capture physical properties of the human body and the forces it experiences. Experiments demonstrate that, once trained, PhysPT can be directly applied to kinematics-based estimates to significantly enhance their physical plausibility and generate favourable motion forces. Furthermore, we show that these physically meaningful quantities translate into improved accuracy of an important downstream task: human action recognition.

Understanding human motion from video is essential for a range of applications, including pose estimation, mesh recovery and action recognition. While state-of-the-art methods predominantly rely on transformer-based architectures, these approaches have limitations in practical scenarios. Transformers are slower when sequentially predicting on a continuous stream of frames in real-time, and do not generalize to new frame rates. In light of these constraints, we propose a novel attention-free spatiotemporal model for human motion understanding building upon recent advancements in state space models.Our model not only matches the performance of transformer-based models in various motion understanding tasks but also brings added benefits like adaptability to different video frame rates and enhanced training speed when working with longer sequence of keypoints. Moreover, the proposed model supports both offline and real-time applications. For real-time sequential prediction, our model is both memory efficient and several times faster than transformer-based approaches while maintaining their high accuracy.

We introduce MultiPhys, a method designed for recovering multi-person motion from monocular videos. Our focus lies in capturing coherent spatial placement between pairs of individuals across varying degrees of engagement. MultiPhys, being physically aware, exhibits robustness to jittering and occlusions, and effectively eliminates penetration issues between the two individuals. We devise a pipeline in which the motion estimated by a kinematic-based method is fed into a physics simulator in an autoregressive manner. We introduce distinct components that enable our model to harness the simulator’s properties without compromising the accuracy of the kinematic estimates. This results in final motion estimates that are both kinematically coherent and physically compliant. Extensive evaluations on three challenging datasets characterized by substantial inter-person interaction show that our method significantly reduces errors associated with penetration and foot skating, while performing competitively with the state-of-the-art on motion accuracy and smoothness. Results and code can be found on our project page.

The credibility and practicality of a reconstructed hand-object interaction sequence depend largely on its physical plausibility. However, due to high occlusions during hand-object interaction, physical plausibility remains a challenging criterion for purely vision-based tracking methods. To address this issue and enhance the results of existing hand trackers, this paper proposes a novel physically-aware hand motion de-noising method. Specifically, we introduce two learned loss terms that explicitly capture two crucial aspects of physical plausibility: grasp credibility and manipulation feasibility. These terms are used to train a physically-aware de-noising network. Qualitative and quantitative experiments demonstrate that our approach significantly improves both fine-grained physical plausibility and overall pose accuracy, surpassing current state-of-the-art de-noising methods.

We address the challenging task of identifying, segmenting, and tracking hand-held objects, which is crucial for applications such as human action segmentation and performance evaluation. This task is particularly challenging due to heavy occlusion, rapid motion, and the transitory nature of objects being hand-held, where an object may be held, released, and subsequently picked up again. To tackle these challenges, we have developed a novel transformer-based architecture called HOIST-Former. HOIST-Former is adept at spatially and temporally segmenting hands and objects by iteratively pooling features from each other, ensuring that the processes of identification, segmentation, and tracking of hand-held objects depend on the hands’ positions and their contextual appearance. We further refine HOIST-Former with a contact loss that focuses on areas where hands are in contact with objects. Moreover, we also contribute an in-the-wild video dataset called HOIST, which comprises 4,125 videos complete with bounding boxes, segmentation masks, and tracking IDs for hand-held objects. Through experiments on the HOIST dataset and two additional public datasets, we demonstrate the efficacy of HOIST-Former in segmenting and tracking hand-held objects.

We present SCULPT, a novel 3D generative model for clothed and textured 3D meshes of humans. Specifically, we devise a deep neural network that learns to represent the geometry and appearance distribution of clothed human bodies. Training such a model is challenging, as datasets of textured 3D meshes for humans are limited in size and accessibility. Our key observation is that there exist medium-sized 3D scan datasets like CAPE, as well as large-scale 2D image datasets of clothed humans and multiple appearances can be mapped to a single geometry. To effectively learn from the two data modalities, we propose an unpaired learning procedure for pose-dependent clothed and textured human meshes. Specifically, we learn a pose-dependent geometry space from 3D scan data. We represent this as per vertex displacements w.r.t. the SMPL model. Next, we train a geometry conditioned texture generator in an unsupervised way using the 2D image data. We use intermediate activations of the learned geometry model to condition our texture generator. To alleviate entanglement between pose and clothing type, and pose and clothing appearance, we condition both the texture and geometry generators with attribute labels such as clothing types for the geometry, and clothing colors for the texture generator. We automatically generated these conditioning labels for the 2D images based on the visual question answering model BLIP and CLIP. We validate our method on the SCULPT dataset, and compare to state-of-the-art 3D generative models for clothed human bodies. We will release the codebase for research purposes.

Recent developments in face restoration have achieved remarkable results in producing high-quality and lifelike outputs. The stunning results however often fail to be faithful with respect to the identity of the person as the models lack necessary context. In this paper, we explore the potential of personalized face restoration with diffusion models. In our approach a restoration model is personalized using a few images of the identity, leading to tailored restoration with respect to the identity while retaining fine-grained details. By using independent trainable blocks for personalization, the rich prior of a base restoration model can be exploited to its fullest. To avoid the model relying on parts of identity left in the conditioning low-quality images, a generative regularizer is employed. With a learnable parameter, the model learns to balance between the details generated based on the input image and the degree of personalization. Moreover, we improve the training pipeline of face restoration models to enable an alignment-free approach. We showcase the robust capabilities of our approach in several real-world scenarios with multiple identities, demonstrating our method's ability to generate fine-grained details with faithful restoration. In the user study we evaluate the perceptual quality and faithfulness of the generated details, with our method being voted best 61\% of the time compared to the second best with 25\% of the votes.

This work proposes a novel learning framework for visual hand dynamics analysis that takes into account the physiological aspects of hand motion. The existing models, which are simplified joint-actuated systems, often produce unnatural motions. To address this, we integrate a musculoskeletal system with a learnable parametric hand model, MANO, to create a new model, MS-MANO. This model emulates the dynamics of muscles and tendons to drive the skeletal system, imposing physiologically realistic constraints on the resulting torque trajectories. We further propose a simulation-in-the-loop pose refinement framework, BioPR, that refines the initial estimated pose through a multi-layer perceptron (MLP) network. Our evaluation of the accuracy of MS-MANO and the efficacy of the BioPR is conducted in two separate parts. The accuracy of MS-MANO is compared with MyoSuite, while the efficacy of BioPR is benchmarked against two large-scale public datasets and two recent state-of-the-art methods. The results demonstrate that our approach consistently improves the baseline methods both quantitatively and qualitatively. Code and models will be made publicly available.

The recently emerging text-to-motion advances have spired numerous attempts for convenient and interactive human motion generation. Yet, existing methods are largely limited to generating body motions only without considering the rich two-hand motions, let alone handling various conditions like body dynamics or texts. To break the data bottleneck, we propose BOTH57M, a novel multi-modal dataset for two-hand motion generation. Our dataset includes accurate motion tracking for the human body and hands and provides pair-wised finger-level hand annotations and body descriptions. We further provide a strong baseline method, BOTH2Hands, for the novel task: generating vivid two-hand motions from both implicit body dynamics and explicit text prompts. We first warm up two parallel body-to-hand and text-to-hand diffusion models and then utilize the cross-attention transformer for motion blending. Extensive experiments and cross-validations demonstrate the effectiveness of our approach and dataset for generating convincing two-hand motions from the hybrid body-and-textual conditions. Our dataset and code will be disseminated to the community for future research, whichcan be found at https://github.com/Godheritage/BOTH2Hands.

DensePose provides a pixel-accurate association of images with 3D mesh coordinates, but does not provide a 3D mesh, while Human Mesh Reconstruction (HMR) systems have high 2D reprojection error, as measured by DensePose localization metrics. In this work we introduce MeshPose to jointly tackle DensePose and HMR. For this we first introduce new losses that allow us to use weak DensePose supervision to accurately localize in 2D a subset of the mesh vertices (‘VertexPose’). We then lift these vertices to 3D, yielding a low-poly body mesh (‘MeshPose’). Our system is trained in an end-to-end manner and is the first HMR method to attain competitive DensePose accuracy, while also being lightweight and amenable to efficient inference, making it suitable for real-time AR applications.

Listening head generation aims to synthesize a non-verbal responsive listener head by modeling the correlation between the speaker and the listener in dynamic conversion. The applications of listener agent generation in virtual interaction have promoted many works achieving the diverse and fine-grained motion generation. However, they can only manipulate motions through simple emotional labels, but cannot freely control the listener's motions. Since listener agents should have human-like attributes (e.g. identity, personality) which can be freely customized by users, this limits their realism. In this paper, we propose a user-friendly framework called CustomListener to realize the free-form text prior guided listener generation. To achieve speaker-listener coordination, we design a Static to Dynamic Portrait module (SDP), which interacts with speaker information to transform static text into dynamic portrait token with completion rhythm and amplitude information. To achieve coherence between segments, we design a Past Guided Generation module (PGG) to maintain the consistency of customized listener attributes through the motion prior, and utilize a diffusion-based structure conditioned on the portrait token and the motion prior to realize the controllable generation. To train and evaluate our model, we have constructed two text-annotated listening head datasets based on ViCo and RealTalk, which provide text-video paired labels. Extensive experiments have verified the effectiveness of our model.

As a significant step for human face modeling, editing, and generation, face landmarking aims at extracting facial keypoints from images. Currently, a generalizable face landmarker is required in practical applications because real-world facial images, e.g., the avatars in animations and games, are often stylized in various ways. However, achieving generalizable face landmarking is often challenging due to the diversity of facial styles and the scarcity of labeled stylized faces. In this study, we propose a simple but effective paradigm for learning a generalizable face landmarker based on labeled real human faces and unlabeled stylized faces. In particular, we learn the face landmarker as the key module of a conditional face warper. Given a pair of real and stylized facial images, the conditional face warper predicts a warping field from the real face to the stylized one, in which the face landmarker predicts the ending points of the warping field and thus provides us with high-quality pseudo landmarks for the corresponding stylized facial images. Applying an alternating optimization strategy, we learn the face landmarker to minimize $i)$ the discrepancy between the stylized faces and the warped real ones and $ii)$ the prediction errors of both real and pseudo landmarks. Extensive experiments on various datasets show that our method outperforms existing state-of-the-art domain adaptation methods in face landmarking tasks, leading to a face landmarker with better generalizability.

In-context learning provides a new perspective for multi-task modeling for vision and NLP. Under this setting, the model can perceive tasks from prompts and accomplish them without any extra task-specific head predictions or model finetuning. However, Skeleton sequence modeling via in-context learning remains unexplored. Directly applying existing in-context models from other areas onto skeleton sequences fails due to the inter-frame and cross-task pose similarity that makes it outstandingly hard to perceive the task correctly from a subtle context. To address this challenge, we propose Skeleton-in-Context (SiC), an effective framework for in-context skeleton sequence modeling. Our SiC is able to handle multiple skeleton-based tasks simultaneously after a single training process and accomplish each task from context according to the given prompt. It can further generalize to new, unseen tasks according to customized prompts. To facilitate context perception, we additionally propose a task-unified prompt, which adaptively learns tasks of different natures, such as partial joint-level generation, sequence-level prediction, or 2D-to-3D motion prediction. We conduct extensive experiments to evaluate the effectiveness of our SiC on multiple tasks, including motion prediction, pose estimation, joint completion, and future pose estimation. We also evaluate its generalization capability on unseen tasks such as motion-in-between. These experiments show that our model achieves state-of-the-art multi-task performance and even outperforms single-task methods on certain tasks.

Canonical emotions, such as happy, sad, and fear, are easy to understand and annotate. However, emotions are often compound, e.g. happily surprised, and can be mapped to the action units (AUs) used for expressing emotions, and trivially to the canonical ones. Intuitively, emotions are continuous as represented by the arousal-valence (AV) model. An interpretable unification of these four modalities —namely, Canonical, Compound, AUs, and AV— is highly desirable, for a better representation and understanding of emotions. However, such unification remains to be unknown in the current literature. In this work, we propose an interpretable and unified emotion model, referred as C2A2. We also develop a method that leverages labels of the non-unified models to annotate the novel unified one. Finally, we modify the text-conditional diffusion models to understand continuous numbers, which are then used to generate continuous expressions using our unified emotion model. Through quantitative and qualitative experiments, we show that our generated images are rich and capture subtle expressions. Our work allows a fine-grained generation of expressions in conjunction with other textual inputs and offers a new label space for emotions at the same time.

An increasingly common approach for creating photo-realistic digital avatars is through the use of volumetric neural fields. The original neural radiance field (NeRF) allowed for impressive novel view synthesis of static heads when trained on a set of multi-view images, and follow up methods showed that these neural representations can be extended to dynamic avatars. Recently, new variants also surpassed the usual drawback of baked-in illumination in neural representations, showing that static neural avatars can be relit in any environment. In this work we simultaneously tackle both the motion and illumination problem, proposing a new method for relightable and animatable neural heads. Our method builds on a proven dynamic avatar approach based on a mixture of volumetric primitives, combined with a recently-proposed lightweight hardware setup for relightable neural fields, and includes a novel architecture that allows relighting dynamic neural avatars performing unseen expressions in any environment, even with nearfield illumination and viewpoints.

Text-to-image generative models can generate high-quality humans, but realism is lost when generating hands. Common artifacts include irregular hand poses, shapes, incorrect numbers of fingers, and physically implausible finger orientations. To generate images with realistic hands, we propose a novel diffusion-based architecture called HanDiffuser that achieves realism by injecting hand embeddings in the generative process. HanDiffuser consists of two components: a Text-to-Hand-Params diffusion model to generate SMPL-Body and MANO-Hand parameters from input text prompts, and a Text-Guided Hand-Params-to-Image diffusion model to synthesize images by conditioning on the prompts and hand parameters generated by the previous component. We incorporate multiple aspects of hand representation, including 3D shapes and joint-level finger positions, orientations and articulations, for robust learning and reliable performance during inference. We conduct extensive quantitative and qualitative experiments and perform user studies to demonstrate the efficacy of our method in generating images with high-quality hands.

Accurately predicting the 3D human posture and the pressure exerted on the body for people resting in bed, visualized as a body mesh (3D pose & shape) with a 3D pressure map, holds significant promise for healthcare applications, particularly, in the prevention of pressure ulcers. Current methods focus on singular facets of the problem---predicting only 2D/3D poses, generating 2D pressure images, predicting pressure only for certain body regions instead of the full body, or forming indirect approximations to the 3D pressure map. In contrast, we introduce BodyMAP, which jointly predicts the human body mesh and 3D applied pressure map across the entire human body. Our network leverages multiple visual modalities, incorporating both a depth image of a person in bed and its corresponding 2D pressure image acquired from a pressure-sensing mattress. The 3D pressure map is represented as a pressure value at each mesh vertex and thus allows for precise localization of high-pressure regions on the body. Additionally, we present BodyMAP-WS, a new formulation of pressure prediction in which we implicitly learn pressure in 3D by aligning sensed 2D pressure images with a differentiable 2D projection of the predicted 3D pressure maps. In evaluations with real-world human data, our method outperforms the current state-of-the-art technique by 25% on both body mesh and 3D applied pressure map prediction tasks for people in bed.

While existing methods to reconstruct 3D faces from monocular in-the-wild images excel at recovering the overall face shape. However, they commonly miss subtle, extreme, and asymmetric expressions, and other rarely observed expressions. We improve upon these methods with SMIRK (Spatial Modeling for Image-based Reconstruction of Kinesics), which faithfully reconstructs expressive 3D faces from images. We identify two key limitations in existing methods that hinder better expression recovery: shortcomings in their self-supervised training formulation, and a lack of expression diversity in the training images. For training, most existing methods employ differentiable rendering to render the predicted face mesh, and compare it to the input image, along with a plethora of additional loss functions. This differentiable rendering loss not only has to provide supervision to optimize for 3D face geometry, camera, albedo, and lighting, which is an ill-posed optimization problem but the domain gap between rendering and input image further hinders the learning process. Instead, SMIRK replaces the differentiable rendering with a neural rendering module that, given the rendered predicted mesh geometry, and sparsely sampled pixels of the input image, generates a face image. As the neural rendering gets color information from sampled image pixels, supervising with neural rendering-based reconstruction loss can focus solely on the geometry. Further, it enables us to generate images of the input identity with varying expressions while training. These are then utilized as input to the reconstruction model and used as supervision with ground truth geometry. This effectively augments the training data and enhances the generalization for diverse expressions. Our qualitative, quantitative and particularly our perceptual evaluations demonstrate that SMIRK achieves the new state-of-the art performance on accurate expression reconstruction.

We present a new self-supervised approach, SelfPose3d, for estimating 3d poses of multiple persons from multiple camera views. Unlike current state-of-the-art fully-supervised methods, our approach does not require any 2d or 3d ground-truth poses and uses only the multi-view input images from a calibrated camera setup and 2d pseudo poses generated from an off-the-shelf 2d human pose estimator. We propose two self-supervised learning objectives: self-supervised person localization in 3d space and self-supervised 3d pose estimation. We achieve self-supervised 3d person localization by training the model on synthetically generated 3d points, serving as 3d person root positions, and on the projected root-heatmaps in all the views. We then model the 3d poses of all the localized persons with a bottleneck representation, map them onto all views obtaining 2d joints, and render them using 2d Gaussian heatmaps in an end-to-end differentiable manner. Afterwards, we use the corresponding 2d joints and heatmaps from the pseudo 2d poses for learning. To alleviate the intrinsic inaccuracy of the pseudo labels, we propose an adaptive supervision attention mechanism to guide the self-supervision. Our experiments and analysis on three public benchmark datasets, including Panoptic, Shelf, and Campus, show the effectiveness of our approach, which is comparable to fully-supervised methods. Code is available at https://github.com/CAMMA-public/SelfPose3D.

Motion capture from a limited number of body-worn sensors, such as inertial measurement units (IMUs) and pressure insoles, has important applications in health, human performance, and entertainment. Recent work has focused on accurately reconstructing whole-body motion from a specific sensor configuration using six IMUs. While a common goal across applications is to use the minimal number of sensors to achieve required accuracy, the optimal arrangement of the sensors might differ from application to application.We propose a single diffusion model, DiffusionPoser, which reconstructs human motion in real-time from arbitrary sensor configurations including IMUs and pressure insoles. Unlike existing methods, our model grants users the flexibility to determine the number and arrangement of sensors tailored to the specific activity of interest, without the need for retraining. A novel autoregressive inferencing scheme ensures real-time motion reconstruction that closely aligns with measured sensor signals. The generative nature of DiffusionPoser ensures realistic behavior, even for degrees-of-freedom not directly measured. Qualitative results can be found on our project website.

Low Light Enhancement (LLE) is an important step to enhance images captured with insufficient light. Several local and global methods have been proposed over the years for this problem. Decomposing the image into multiple factors using an appropriate property is the first step in many LLE methods. In this paper, we present a new additive factorization that treats images to be composed of multiple latent specular components that can be estimated by modulating the sparsity during decomposition. We propose a model-driven learnable RSFNet framework to estimate these factors by unrolling the optimization into network layers. The factors are interpretable by design and can be manipulated directly for different tasks. We train our LLE system in a {\em zero-reference} manner without the need for any paired or unpaired supervision. Our system improves the state-of-the-art performance on standard benchmarks and achieves better generalization on multiple other datasets. The specularity factors can supplement other task specific fusion networks by inducing prior information for enhancement tasks like deraining, deblurring and dehazing with negligible overhead as shown in the paper.

Diffusion Models have shown remarkable performance in image generation tasks, which are capable of generating diverse and realistic image content. When adopting diffusion models for image restoration, the crucial challenge lies in how to preserve high-level image fidelity in the randomness diffusion process and generate accurate background structures and realistic texture details. In this paper, we propose a general framework and develop a Diffusion Texture Prior Model (DTPM) for image restoration tasks. DTPM explicitly models high-quality texture details through the diffusion process, rather than global contextual content. In phase one of the training stage, we pre-train DTPM on approximately 55K high-quality image samples, after which we freeze most of its parameters. In phase two, we insert conditional guidance adapters into DTPM and equip it with an initial predictor, thereby facilitating its rapid adaptation to downstream image restoration tasks. Our DTPM could mitigate the randomness of traditional diffusion models by utilizing encapsulated rich and diverse texture knowledge and background structural information provided by the initial predictor during the sampling process. Our comprehensive evaluations of five image restoration tasks demonstrate DTPM's superiority over existing regression and diffusion-based image restoration methods in perceptual quality and its exceptional generalization capabilities. The code will be publicly available.

Text-based diffusion models have exhibited remarkable success in generation and editing, showing great promise for enhancing visual content with their generative prior. However, applying these models to video super-resolution remains challenging due to the high demands for output fidelity and temporal consistency, which is complicated by the inherent randomness in diffusion models. Our study introduces Upscale-A-Video, a text-guided latent diffusion framework for video upscaling. This framework ensures temporal coherence through two key mechanisms: locally, it integrates temporal layers into U-Net and VAE-Decoder, maintaining consistency in short sequences; globally, without training, a flow-guided latent propagation module is introduced to enhance overall video stability by propagating and fusing latent across the entire sequences. Thanks to the diffusion paradigm, our model also offers greater flexibility by allowing text prompts to guide texture creation and adjustable noise levels to balance restoration and generation, enabling a trade-off between fidelity and quality. Extensive experiments show that Upscale-A-Video surpasses existing methods in both synthetic and real-world benchmarks, as well as in AI-generated videos, showcasing impressive visual realism and temporal consistency.

In video super-resolution, it is common to use a frame-wise alignment to support the propagation of information over time. The role of alignment is well-studied for low-level enhancement in video, but existing works overlook a critical step -- resampling. We show through extensive experiments that for alignment to be effective, the resampling should preserve the reference frequency spectrum while minimizing spatial distortions. However, most existing works simply use a default choice of bilinear interpolation for resampling even though bilinear interpolation has a smoothing effect and hinders super-resolution. From these observations, we propose an implicit resampling-based alignment. The sampling positions are encoded by a sinusoidal positional encoding, while the value is estimated with a coordinate network and a window-based cross-attention. We show that bilinear interpolation inherently attenuates high-frequency information while an MLP-based coordinate network can approximate more frequencies. Experiments on synthetic and real-world datasets show that alignment with our proposed implicit resampling enhances the performance of state-of-the-art frameworks with minimal impact on both compute and parameters.

Implicit neural representations (INRs) have emerged as a promising approach for video storage and processing, showing remarkable versatility across various video tasks. However, existing methods often fail to fully leverage their representation capabilities, primarily due to inadequate alignment of intermediate features during target frame decoding. This paper introduces a universal boosting framework for current implicit video representation approaches. Specifically, we utilize a conditional decoder with a temporal-aware affine transform module, which uses the frame index as a prior condition to effectively align intermediate features with target frames. Besides, we introduce a sinusoidal NeRV-like block to generate diverse intermediate features and achieve a more balanced parameter distribution, thereby enhancing the model's capacity. With a high-frequency information-preserving reconstruction loss, our approach successfully boosts multiple baseline INRs in the reconstruction quality and convergence speed for video regression, and exhibits superior inpainting and interpolation results. Further, we integrate a consistent entropy minimization technique and develop video codecs based on these boosted INRs. Experiments on the UVG dataset confirm that our enhanced codecs significantly outperform baseline INRs and offer competitive rate-distortion performance compared to traditional and learning-based codecs.

Image enhancement holds extensive applications in real-world scenarios due to complex environments and limitations of imaging devices. Conventional methods are often constrained by their tailored models, resulting in diminished robustness when confronted with challenging degradation conditions. In response, we propose FlowIE, a simple yet highly effective flow-based image enhancement framework that estimates straight-line paths from an elementary distribution to high-quality images. Unlike previous diffusion-based methods that suffer from long-time inference, FlowIE constructs a linear many-to-one transport mapping via conditioned rectified flow. The rectification straightens the trajectories of probability transfer, accelerating inference by an order of magnitude. This design enables our FlowIE to fully exploit rich knowledge in the pre-trained diffusion model, rendering it well-suited for various real-world applications. Moreover, we devise a faster inference algorithm, inspired by Lagrange's Mean Value Theorem, harnessing midpoint tangent direction to optimize path estimation, ultimately yielding visually superior results. Thanks to these designs, our FlowIE adeptly manages a diverse range of enhancement tasks within a concise sequence of fewer than 5 steps. Our contributions are rigorously validated through comprehensive experiments on synthetic and real-world datasets, unveiling the compelling efficacy and efficiency of our proposed FlowIE. Code is available at \url{https://github.com/EternalEvan/FlowIE}.

The inherent generative power of denoising diffusion models makes them well-suited for image restoration tasks where the objective is to find the optimal high-quality image within the generative space that closely resembles the input image.We propose a method to adapt a pretrained diffusion model for image restoration by simply adding noise to the input image to be restored and then denoise. Our method is based on the observation that the space of a generative model needs to be constrained. We impose this constraint by finetuning the generative model with a set of anchor images that capture the characteristics of the input image. With the constrained space, we can then leverage the sampling strategy used for generation to do image restoration. We evaluate against previous methods and show superior performances on multiple real-world restoration datasets in preserving identity and image quality. We also demonstrate an important and practical application on personalized restoration, where we use a personal album as the anchor images to constrain the generative space. This approach allows us to produce results that accurately preserve high-frequency details, which previous works are unable to do.

Event camera has recently received much attention for low-light image enhancement (LIE) thanks to their distinct advantages, such as high dynamic range. However, current research is prohibitively restricted by the lack of large-scale, real-world, and spatial-temporally aligned event-image datasets. To this end, we propose a real-world (indoor and outdoor) dataset comprising over 30K pairs of images and events under both low and normal illumination conditions. To achieve this, we utilize a robotic arm that traces a consistent non-linear trajectory to curate the dataset with spatial alignment precision under 0.03mm. We then introduce a matching alignment strategy, rendering 90% of our dataset with errors less than 0.01s. Based on the dataset, we propose a novel event-guided LIE approach, called EvLight, towards robust performance in real-world low-light scenes. Specifically, we first design the multi-scale holistic fusion branch to extract holistic structural and textural information from both events and images. To ensure robustness against variations in the regional illumination and noise, we then introduce a Signal-to-Noise-Ratio (SNR)-guided regional feature selection to selectively fuse features of images from regions with high SNR and enhance those with low SNR by extracting regional structural information from events. our EvLight significantly surpasses the frame-based methods, e.g., Retinexformer by 1.14 dB and 2.62 dB, respectively. Code and datasets are available at https://vlislab22.github.io/eg-lowlight/.

Event Stream Super-Resolution (ESR) aims to address the challenge of insufficient spatial resolution in event streams, which holds great significance for the application of event cameras in complex scenarios. Previous works for ESR often process positive and negative events in a mixed paradigm. This paradigm limits their ability to effectively model the unique characteristics of each event and mutually refine each other by considering their correlations. In this paper, we propose a bilateral event mining and complementary network (BMCNet) to fully leverage the potential of each event and capture the shared information to complement each other simultaneously. Specifically, we resort to a two-stream network to accomplish comprehensive mining of each type of events individually. To facilitate the exchange of information between two streams, we propose a bilateral information exchange (BIE) module. This module is layer-wisely embedded between two streams, enabling the effective propagation of hierarchical global information while alleviating the impact of invalid information brought by inherent characteristics of events. The experimental results demonstrate that our approach outperforms the previous state-of-the-art methods in ESR, achieving performance improvements of over 11% on both real and synthetic datasets. Moreover, our method significantly enhances the performance of event-based downstream tasks such as object recognition and video reconstruction.

As an important pillar of underwater intelligence, Marine Animal Segmentation (MAS) involves segmenting animals within marine environments. Previous methods don’t excel in extracting long-range contextual features and overlook the connectivity between discrete pixels. Recently, Segment Anything Model (SAM) offers a universal framework for general segmentation tasks. Unfortunately, trained with natural images, SAM does not obtain the prior knowledge from marine images. In addition, the single-position prompt of SAM is very insufficient for prior guidance. To address these issues, we propose a novel feature learning framework, named Dual-SAM for high-performance MAS. To this end, we first introduce a dual structure with SAM’s paradigm to enhance feature learning of marine images. Then, we propose a Multi-level Coupled Prompt (MCP) strategy to instruct comprehensive underwater prior information, and enhance the multi-level features of SAM’s encoder with adapters. Subsequently, we design a Dilated Fusion Attention Module (DFAM) to progressively integrate multi-level features from SAM’s encoder. Finally, instead of directly predicting the masks of marine animals, we propose a Criss-Cross Connectivity Prediction ($C^3$P) paradigm to capture the inter-connectivity between discrete pixels. With dual decoders, it generates pseudo-labels and achieves mutual supervision for complementary feature representations, resulting in considerable improvements over previous techniques. Extensive experiments verify that our proposed method achieves state-of-the-art performances on five widely-used MAS datasets. The code is available at https://github.com/Drchip61/Dual_SAM.

Estimating large, extreme inter-image rotations is critical for numerous computer vision domains involving images related by limited or non-overlapping fields of view. In this work, we propose an attention-based approach with a pipeline of novel algorithmic components. First, as rotation estimation pertains to image pairs, we introduce an inter-image distillation scheme using Decoders to improve embeddings. Second, whereas contemporary methods compute a 4D correlation volume (4DCV) encoding inter-image relationships, we propose an Encoder-based cross-attention approach between activation maps to compute an enhanced equivalent of the 4DCV. Finally, we present a cascaded Decoder-based technique for alternately refining both the cross-attention and the rotation query. Our approach outperforms current state-of-the-art methods on extreme rotation estimation. We make our code publicly available.

Panoramic videos have the advantage of providing an immersive and interactive viewing experience. Nevertheless, their spherical nature gives rise to various and uncertain user viewing behaviors, which poses significant challenges for panoramic video quality assessment (PVQA). In this work, we propose an end-to-end optimized, blind PVQA method with explicit modeling of user viewing patterns through visual scanpaths. Our method consists of two modules: a scanpath generator and a quality assessor. The scanpath generator is initially trained to predict future scanpaths by minimizing their expected code length and then jointly optimized with the quality assessor for quality prediction. Our blind PVQA method enables direct quality assessment of panoramic images by treating them as videos composed of identical frames. Experiments on three public panoramic image and video quality datasets, encompassing both synthetic and authentic distortions, validate the superiority of our blind PVQA model over existing methods.

We propose a framework for automatic colorization that allows for iterative editing and modifications. The core of our framework lies in an imagination module: by understanding the content within a grayscale image, we utilize a pre-trained image generation model to generate multiple images that contain the same content. These images serve as references for coloring, mimicking the process of human experts. As the synthesized images can be imperfect or different from the original grayscale image, we propose a Reference Refinement Module to select the optimal reference composition. Unlike most previous end-to-end automatic colorization algorithms, our framework allows for iterative and localized modifications of the colorization results because we explicitly model the coloring samples. Extensive experiments demonstrate the superiority of our framework over existing automatic colorization algorithms in editability and flexibility. Project page: https://xy-cong.github.io/imagine-colorization/.

Scene-aware Adaptive Compressive Sensing (ACS) has constituted a persistent pursuit, holding substantial promise for the enhancement of Compressive Sensing (CS) performance. Cascaded ACS furnishes a proficient multi-stage framework for adaptively allocating the CS sampling based on previous CS measurements. However, reconstruction is commonly required for analyzing and steering the successive CS sampling, which bottlenecks the ACS speed and impedes the practical application in time-sensitive scenarios.Addressing this challenge, we propose a reconstruction-free cascaded ACS method, which requires NO reconstruction during the adaptive sampling process. A lightweight Score Network (ScoreNet) is proposed to directly determine the ACS allocation with previous CS measurements and a differentiable adaptive sampling module is proposed for end-to-end training. For image reconstruction, we propose a Multi-Grid Spatial-Attention Network (MGSANet) that could facilitate efficient multi-stage training and inferencing. By introducing the reconstruction-fidelity supervision outside the loop of the multi-stage sampling process, ACS can be efficiently optimized and achieve high imaging fidelity. The effectiveness of the proposed method is demonstrated with extensive quantitative and qualitative experiments, compared with the state-of-the-art CS algorithms.

Existing research based on deep learning has extensively explored the problem of daytime image dehazing. However, few studies have considered the characteristics of nighttime hazy scenes. There are two distinctions between nighttime and daytime haze. First, there may be multiple active colored light sources with lower illumination intensity in nighttime scenes, which may cause haze, glow and noise with localized, coupled and frequency inconsistent characteristics. Second, due to the domain discrepancy between simulated and real-world data, unrealistic brightness may occur when applying a dehazing model trained on simulated data to real-world data. To address the above two issues, we propose a semi-supervised model for real-world nighttime dehazing. First, the spatial attention and frequency spectrum filtering are implemented as a spatial-frequency domain information interaction module to handle the first issue. Second, a pseudo-label-based retraining strategy and a local window-based brightness loss for semi-supervised training process is designed to suppress haze and glow while achieving realistic brightness. Experiments on public benchmarks validate the effectiveness of the proposed method and its superiority over state-of-the-art methods. The source code and Supplementary Materials are placed in the https://github.com/Xiaofeng-life/SFSNiD.

Although image super-resolution (SR) problem has experienced unprecedented restoration accuracy with deep neural networks, it has yet limited versatile applications due to the substantial computational costs. Since different input images for SR face different restoration difficulties, adapting computational costs based on the input image, referred to as adaptive inference, has emerged as a promising solution to compress SR networks. Specifically, adapting the quantization bit-widths has successfully reduced the inference and memory cost without sacrificing the accuracy. However, despite the benefits of the resultant adaptive network, existing works rely on time-intensive quantization-aware training with full access to the original training pairs to learn the appropriate bit allocation policies, which limits its ubiquitous usage. To this end, we introduce the first on-the-fly adaptive quantization framework that accelerates the processing time from hours to seconds. We formulate the bit allocation problem with only two bit mapping modules: one to map the input image to the image-wise bit adaptation factor and one to obtain the layer-wise adaptation factors. These bit mappings are calibrated and fine-tuned using only a small number of calibration images. We achieve competitive performance with the previous adaptive quantization methods, while the processing time is accelerated by $\times$2000. Codes are available at https://github.com/Cheeun/AdaBM.

In real-world scenarios, image recognition tasks, such as semantic segmentation and object detection, often pose greater challenges due to the lack of information available within low-resolution (LR) content. Image super-resolution (SR) is one of the promising solutions for addressing the challenges. However, due to the ill-posed property of SR, it is challenging for typical SR methods to restore task-relevant high-frequency contents, which may dilute the advantage of utilizing the SR method. Therefore, in this paper, we propose Super-Resolution for Image Recognition (SR4IR) that effectively guides the generation of SR images beneficial to achieving satisfactory image recognition performance when processing LR images. The critical component of our SR4IR is the task-driven perceptual (TDP) loss that enables the SR network to acquire task-specific knowledge from a network tailored for a specific task. Moreover, we propose a cross-quality patch mix and an alternate training framework that significantly enhances the efficacy of the TDP loss by addressing potential problems when employing the TDP loss. Through extensive experiments, we demonstrate that our SR4IR achieves outstanding task performance by generating SR images useful for a specific image recognition task, including semantic segmentation, object detection, and image classification. The implementation code is available at https://github.com/JaehaKim97/SR4IR.

Image Quality Assessment (IQA) constitutes a fundamental task within the field of computer vision, yet it remains an unresolved challenge, owing to the intricate distortion conditions, diverse image contents, and limited availability of data. Recently, the community has witnessed the emergence of numerous large-scale pretrained foundation models. However, it remains an open problem whether the scaling law in high-level tasks is also applicable to IQA tasks which are closely related to low-level clues. In this paper, we demonstrate that with a proper injection of local distortion features, a larger pretrained vision transformer (ViT) foundation model performs better in IQA tasks. Specifically, for the lack of local distortion structure and inductive bias of the large-scale pretrained ViT, we use another pretrained convolution neural networks (CNNs), which is well known for capturing the local structure, to extract multi-scale image features. Further, we propose a local distortion extractor to obtain local distortion features from the pretrained CNNs and a local distortion injector to inject the local distortion features into ViT. By only training the extractor and injector, our method can benefit from the rich knowledge in the powerful foundation models and achieve state-of-the-art performance on popular IQA datasets, indicating that IQA is not only a low-level problem but also benefits from stronger high-level features drawn from large-scale pretrained models. Codes are publicly available at: https://github.com/NeosXu/LoDa.

Video deblurring relies on leveraging information from other frames in the video sequence to restore the blurred regions in the current frame. Mainstream approaches employ bidirectional feature propagation, spatio-temporal transformers, or a combination of both to extract information from the video sequence. However, limitations in memory and computational resources constraints the temporal window length of the spatio-temporal transformer, preventing the extraction of longer temporal contextual information from the video sequence. Additionally, bidirectional feature propagation is highly sensitive to inaccurate optical flow in blurry frames, leading to error accumulation during the propagation process. To address these issues, we propose BSSTNet, Blur-aware Spatio-temporal Sparse Transformer Network.It introduces the blur map, which converts the originally dense attention into a sparse form, enabling a more extensive utilization of information throughout the entire video sequence. Specifically, BSSTNet (1) uses a longer temporal window in the transformer, leveraging information from more distant frames to restore the blurry pixels in the current frame. (2) introduces bidirectional feature propagation guided by blur maps, which reduces error accumulation caused by the blur frame. The experimental results demonstrate the proposed BSSTNet outperforms the state-of-the-art methods on the GoPro and DVD datasets.

We introduce a lightweight and accurate architecture for resource-efficient visual correspondence. Our method, dubbed XFeat (Accelerated Features), revisits fundamental design choices in convolutional neural networks for detecting, extracting, and matching local features. Our new model satisfies a critical need for fast and robust algorithms suitable to resource-limited devices. In particular, accurate image matching requires sufficiently large image resolutions -- for this reason, we keep the resolution as large as possible while limiting the number of channels in the network. Besides, our model is designed to offer the choice of matching at the sparse or semi-dense levels, each of which may be more suitable for different downstream applications, such as visual navigation and augmented reality. Our model is the first to offer semi-dense matching efficiently, leveraging a novel match refinement module that relies on coarse local descriptors. XFeat is versatile and hardware-independent, surpassing current deep learning-based local features in speed (up to 5x faster) with comparable or better accuracy, proven in pose estimation and visual localization. We showcase it running in real-time on an inexpensive laptop CPU without specialized hardware optimizations. Code and weights are available at verlab.dcc.ufmg.br/descriptors/xfeat_cvpr24.

Image stitching from different captures often results in non-rectangular boundaries, which is often considered unappealing. To solve non-rectangular boundaries, current solutions involve cropping, which discards image content, inpainting, which can introduce unrelated content, or warping, which can distort non-linear features and introduce artifacts. To overcome these issues, we introduce a novel diffusion-based learning framework, $\textbf{RecDiffusion}$, for image stitching rectangling. This framework combines Motion Diffusion Models (MDM) to generate motion fields, effectively transitioning from the stitched image's irregular borders to a geometrically corrected intermediary. Followed by Content Diffusion Models (CDM) for image detail refinement. Notably, our sampling process utilizes a weighted map to identify regions needing correction during each iteration of CDM. Our RecDiffusion ensures geometric accuracy and overall visual appeal, surpassing all previous methods in both quantitative and qualitative measures when evaluated on public benchmarks. Code is released at https://github.com/lhaippp/RecDiffusion.

The significant amount of manual efforts in annotating pixel-level labels has triggered the advancement of unsupervised saliency learning. However, without supervision signals, state-of-the-art methods can only infer region-level saliency. In this paper, we propose to explore the unsupervised salient instance detection (USID) problem, for a more fine-grained visual understanding. Our key observation is that self-supervised transformer features may exhibit local similarities as well as different levels of contrast to other regions, which provide informative cues to identify salient instances. Hence, we propose CoCo, a novel network that models saliency coherence and contrast for USID. SCoCo includes two novel modules: (1) a global background adaptation (GBA) module with a scene-level contrastive loss to extract salient regions from the scene by searching the adaptive “saliency threshold” in the self-supervised transformer features, and (2) a locality-aware similarity (LAS) module with an instance-level contrastive loss to group salient regions into instances by modeling the in-region saliency coherence and cross-region saliency contrasts. Extensive experiments show that SCoCo outperforms state-of-the-art weakly-supervised SID methods and carefully designed unsupervised baselines, and has comparable performances to fully-supervised SID methods.

Implicit Neural Representation (INR), which utilizes a neural network to map coordinate inputs to corresponding attributes, is causing a revolution in the field of signal processing. However, current INR techniques suffer from a restricted capability to tune their supported frequency set, resulting in imperfect performance when representing complex signals with multiple frequencies. We have identified that this frequency-related problem can be greatly alleviated by introducing variable-periodic activation functions, for which we propose FINER. By initializing the bias of the neural network within different ranges, sub-functions with various frequencies in the variable-periodic function are selected for activation. Consequently, the supported frequency set of FINER can be flexibly tuned, leading to improved performance in signal representation. We demonstrate the capabilities of FINER in the contexts of 2D image fitting, 3D signed distance field representation, and 5D neural radiance fields optimization, and we show that it outperforms existing INRs.

We present a joint learning scheme of video super-resolution and deblurring, called VSRDB, to restore clean high-resolution (HR) videos from blurry low-resolution (LR) ones. This joint restoration problem has drawn much less attention compared to single restoration problems. In this paper, we propose a novel flow-guided dynamic filtering (FGDF) and iterative feature refinement with multi-attention (FRMA), which constitutes our VSRDB framework, denoted as FMA-Net. Specifically, our proposed FGDF enables precise estimation of both spatio-temporally-variant degradation and restoration kernels that are aware of motion trajectories through sophisticated motion representation learning. Compared to conventional dynamic filtering, the FGDF enables the FMA-Net to effectively handle large motions into the VSRDB. Additionally, the stacked FRMA blocks trained with our novel temporal anchor (TA) loss, which temporally anchors and sharpens features, refine features in a coarse-to-fine manner through iterative updates. Extensive experiments demonstrate the superiority of the proposed FMA-Net over state-of-the-art methods in terms of both quantitative and qualitative quality. Codes and pre-trained models are available at: https://kaist-viclab.github.io/fmanet-site.

Image denoising approaches based on deep neural networks often struggle with overfitting to specific noise distributions present in training data. This challenge persists in existing real-world denoising networks, which are trained using a limited spectrum of real noise distributions, and thus, show poor robustness to out-of-distribution real noise types. To alleviate this issue, we develop a novel training framework called Adversarial Frequency Mixup (AFM). AFM leverages mixup in the frequency domain to generate noisy images with distinctive and challenging noise characteristics, all the while preserving the properties of authentic real-world noise. Subsequently, incorporating these noisy images into the training pipeline enhances the denoising network's robustness to variations in noise distributions. Extensive experiments and analyses, conducted on a wide range of real noise benchmarks demonstrate that denoising networks trained with our proposed framework exhibit significant improvements in robustness to unseen noise distributions. Code is available at https://github.com/dhryougit/AFM

Coarse-to-fine schemes are widely used in traditional single-image motion deblur; however, in the context of deep learning, existing multi-scale algorithms not only require the use of complex modules for feature fusion of low-scale RGB images and deep semantics, but also manually generate low-resolution pairs of images that do not have sufficient confidence. In this work, we propose a multi-scale network based on single-input and multiple-outputs(SIMO) for motion deblurring. This simplifies the complexity of algorithms based on a coarse-to-fine scheme. To alleviate restoration defects impacting detail information brought about by using a multi-scale architecture, we combine the characteristics of real-world blurring trajectories with a learnable wavelet transform module to focus on the directional continuity and frequency features of the step-by-step transitions between blurred images to sharp images. In conclusion, we propose a multi-scale network with a learnable discrete wavelet transform (MLWNet), which exhibits state-of-the-art performance on multiple real-world deblurred datasets, in terms of both subjective and objective quality as well as computational efficiency. Our code is available on https://github.com/thqiu0419/MLWNet.

Scene recovery, the restoration of images degraded by adverse weather conditions, presents significant challenges for existing methods. Physical models, constrained by their inherent assumptions, often fail when these assumptions are not met; Deep learning models are powerful, they are limited by the diversity of their training datasets, leading to poor generalization and high computational demands. To address these limitations, we propose the Luminous Flux Prior (LFP), to recover degraded images under diverse adverse weather without learning. Luminous flux, a physical measure that reflects image brightness, has a rate of change that demonstrates a significant correlation with transmission. Consequently, we leverage this rate of change in luminous flux as prior knowledge to estimate transmission, which in turn assists in image recovery. This approach reduces dependency on physical parameters and enhances adaptability to various weather. Experimental validation under diverse conditions, such as sandstorms, underwater environments, and haze, attests to the robustness of LFP in restoring clear images. With a time complexity of $\mathcal{O}(N\log N)$, LFP enables real-time recovery, making it a suitable for devices with limited computational resources.

Previous methods for Video Frame Interpolation (VFI) have encountered challenges, notably the manifestation of blur and ghosting effects. These issues can be traced back to two pivotal factors: unavoidable motion errors and misalignment in supervision. In practice, motion estimates often prove to be error-prone, resulting in misaligned features. Furthermore, the reconstruction loss tends to bring blurry results, particularly in misaligned regions. To mitigate these challenges, we propose a new paradigm called PerVFI (Perception-oriented Video Frame Interpolation). Our approach incorporates an Asymmetric Synergistic Blending module (ASB) that utilizes features from both sides to synergistically blend intermediate features. One reference frame emphasizes primary content, while the other contributes complementary information. To impose a stringent constraint on the blending process, we introduce a self-learned sparse quasi-binary mask which effectively mitigates ghosting and blur artifacts in the output. Additionally, we employ a normalizing flow-based generator and utilize the negative log-likelihood loss to learn the conditional distribution of the output, which further facilitates the generation of clear and fine details. Experimental results validate the superiority of PerVFI, demonstrating significant improvements in perceptual quality compared to existing methods. Codes will be released upon publication.

Blind video quality assessment (BVQA) plays a pivotal role in evaluating and improving the viewing experience of end-users across a wide range of video-based platforms and services. Contemporary deep learning-based models primarily analyze video content in its aggressively subsampled format, while being blind to the impact of the actual spatial resolution and frame rate on video quality. In this paper, we propose a modular BVQA model and a method of training it to improve its modularity. Our model comprises a base quality predictor, a spatial rectifier, and a temporal rectifier, responding to the visual content and distortion, spatial resolution, and frame rate changes on video quality, respectively. During training, spatial and temporal rectifiers are dropped out with some probabilities to render the base quality predictor a standalone BVQA model, which should work better with the rectifiers. Extensive experiments on both professionally-generated content and user-generated content video databases show that our quality model achieves superior or comparable performance to current methods. Additionally, the modularity of our model offers an opportunity to analyze existing video quality databases in terms of their spatial and temporal complexity.

We propose residual denoising diffusion models (RDDM), a novel dual diffusion process that decouples the traditional single denoising diffusion process into residual diffusion and noise diffusion. This dual diffusion framework expands the denoising-based diffusion models, initially uninterpretable for image restoration, into a unified and interpretable model for both image generation and restoration by introducing residuals. Specifically, our residual diffusion represents directional diffusion from the target image to the degraded input image and explicitly guides the reverse generation process for image restoration, while noise diffusion represents random perturbations in the diffusion process. The residual prioritizes certainty, while the noise emphasizes diversity, enabling RDDM to effectively unify tasks with varying certainty or diversity requirements, such as image generation and restoration. We demonstrate that our sampling process is consistent with that of DDPM and DDIM through coefficient transformation, and propose a partially path-independent generation process to better understand the reverse process. Notably, our RDDM enables a generic UNet, trained with only an L1 loss and a batch size of 1, to compete with state-of-the-art image restoration methods. We provide code and pre-trained models to encourage further exploration, application, and development of our innovative framework (https://github.com/nachifur/RDDM).

We propose a practical approach to JPEG image decoding, utilizing a local implicit neural representation with continuous cosine formulation. The JPEG algorithm significantly quantizes discrete cosine transform (DCT) spectra to achieve a high compression rate, inevitably resulting in quality degradation while encoding an image. We have designed a continuous cosine spectrum estimator to address the quality degradation issue that restores the distorted spectrum. By leveraging local DCT formulations, our network has the privilege to exploit dequantization and upsampling simultaneously. Our proposed model enables decoding compressed images directly across different quality factors using a single pre-trained model without relying on a conventional JPEG decoder. As a result, our proposed network achieves state-of-the-art performance in flexible color image JPEG artifact removal tasks. Our source code is available at https://github.com/WooKyoungHan/JDEC

Recent advances in monocular depth estimation have been made by incorporating natural language as additional guidance. Although yielding impressive results, the impact of the language prior, particularly in terms of generalization and robustness, remains unexplored. In this paper, we address this gap by quantifying the impact of this prior and introduce methods to benchmark its effectiveness across various settings. We generate "low-level" sentences that convey object-centric, three-dimensional spatial relationships, incorporate them as additional language priors and evaluate their downstream impact on depth estimation. Our key finding is that current language-guided depth estimators perform optimally only with scene-level descriptions and counter-intuitively fare worse with low level descriptions. Despite leveraging additional data, these methods are not robust to directed adversarial attacks and decline in performance with an increase in distribution shift. Finally, to provide a foundation for future research, we identify points of failures and offer insights to better understand these shortcomings. With an increasing number of methods using language for depth estimation, our findings highlight the opportunities and pitfalls that require careful consideration for effective deployment in real-world settings.

This paper presents an innovative framework designed to train an image deblurring algorithm tailored to a specific camera device. This algorithm works by transforming a blurry input image, which is challenging to deblur, into another blurry image that is more amenable to deblurring. The transformation process, from one blurry state to another, leverages unpaired data consisting of sharp and blurry images captured by the target camera device. Learning this blur-to-blur transformation is inherently simpler than direct blur-to-sharp conversion, as it primarily involves modifying blur patterns rather than the intricate task of reconstructing fine image details. The efficacy of the proposed approach has been demonstrated through comprehensive experiments on various benchmarks, where it significantly outperforms state-of-the-art methods both quantitatively and qualitatively.

Self-supervised denoising has attracted widespread attention due to its ability to train without clean images. However, noise in real-world scenarios is often spatially correlated, which causes many self-supervised algorithms that assume pixel-wise independent noise to perform poorly. Recent works have attempted to break noise correlation with downsampling or neighborhood masking. However, denoising on downsampled subgraphs can lead to aliasing effects and loss of details due to a lower sampling rate. Furthermore, the neighborhood masking methods either come with high computational complexity or do not consider local spatial preservation during inference. Through the analysis of existing methods, we point out that the key to obtaining high-quality and texture-rich results in real-world self-supervised denoising tasks is to train at the original input resolution structure and use asymmetric operations during training and inference. Based on this, we propose Asymmetric Tunable Blind-Spot Network (AT-BSN), where the blind-spot size can be freely adjusted, thus better balancing noise correlation suppression and image local spatial destruction during training and inference. In addition, we regard the pre-trained AT-BSN as a meta-teacher network capable of generating various teacher networks by sampling different blind-spots. We propose a blind-spot based multi-teacher distillation strategy to distill a lightweight network, significantly improving performance. Experimental results on multiple datasets prove that our method achieves state-of-the-art, and is superior to other self-supervised algorithms in terms of computational overhead and visual effects.

Diffusion model-based image restoration (IR) aims to use diffusion models to recover high-quality (HQ) images from degraded images, achieving promising performance. Due to the inherent property of diffusion models, most existing methods need long serial sampling chains to restore HQ images step-by-step, resulting in expensive sampling time and high computation costs. Moreover, such long sampling chains hinder understanding the relationship between inputs and restoration results since it is hard to compute the gradients in the whole chains. In this work, we aim to rethink the diffusion model-based IR models through a different perspective, i.e., a deep equilibrium (DEQ) fixed point system, called DeqIR. Specifically, we derive an analytical solution by modeling the entire sampling chain in these IR models as a joint multivariate fixed point system. Based on the analytical solution, we can conduct parallel sampling and restore HQ images without training. Furthermore, we compute fast gradients via DEQ inversion and found that initialization optimization can boost image quality and control the generation direction. Extensive experiments on benchmarks demonstrate the effectiveness of our method on typical IR tasks and real-world settings.

Video quality assessment (VQA) is a challenging problem due to the numerous factors that can affect the perceptual quality of a video, \eg, content attractiveness, distortion type, motion pattern, and level. However, annotating the Mean opinion score (MOS) for videos is expensive and time-consuming, which limits the scale of VQA datasets, and poses a significant obstacle for deep learning-based methods. In this paper, we propose a VQA method named PTM-VQA, which leverages PreTrained Models to transfer knowledge from models pretrained on various pre-tasks, enabling benefits for VQA from different aspects.Specifically, we extract features of videos from different pretrained models with frozen weights and integrate them to generate representation. Since these models possess various fields of knowledge and are often trained with labels irrelevant to quality, we propose an Intra-Consistency and Inter-Divisibility (ICID) loss to impose constraints on features extracted by multiple pretrained models. The intra-consistency constraint ensures that features extracted by different pretrained models are in the same unified quality-aware latent space, while the inter-divisibility introduces pseudo clusters based on the annotation of samples and tries to separate features of samples from different clusters. Furthermore, with a constantly growing number of pretrained models, it is crucial to determine which models to use and how to use them. To address this problem, we propose an efficient scheme to select suitable candidates. Models with better clustering performance on VQA datasets are chosen to be our candidates. Extensive experiments demonstrate the effectiveness of the proposed method.

Recovering a clear image from a single hazy image is an open inverse problem. Although significant research progress has been made, most existing methods ignore the effect that downstream tasks play in promoting upstream dehazing. From the perspective of the haze generation mechanism, there is a potential relationship between the depth information of the scene and the hazy image. Based on this, we propose a dual-task collaborative mutual promotion framework to achieve the dehazing of a single image. This framework integrates depth estimation and dehazing by a dual-task interaction mechanism and achieves mutual enhancement of their performance. To realize the joint optimization of the two tasks, an alternative implementation mechanism with the difference perception is developed. On the one hand, the difference perception between the depth maps of the dehazing result and the ideal image is proposed to promote the dehazing network to pay attention to the non-ideal areas of the dehazing. On the other hand, by improving the depth estimation performance in the difficult-to-recover areas of the hazy image, the dehazing network can explicitly use the depth information of the hazy image to assist the clear image recovery. To promote the depth estimation, we propose to use the difference between the dehazed image and the ground truth to guide the depth estimation network to focus on the dehazed unideal areas. It allows dehazing and depth estimation to leverage their strengths in a mutually reinforcing manner. Experimental results show that the proposed method can achieve better performance than that of the state-of-the-art approaches.

Single Image Super-Resolution is a classic computer vision problem that involves estimating high-resolution (HR) images from low-resolution (LR) ones. Although deep neural networks (DNNs), especially Transformers for super-resolution, have seen significant advancements in recent years, challenges still remain, particularly in limited receptive field caused by window-based self-attention. To address these issues, we introduce a group of auxiliary Adaptive Token Dictionary to SR Transformer and establish an ATD-SR method. The introduced token dictionary could learn prior information from training data and adapt the learned prior to specific testing image through an adaptive refinement step. The refinement strategy could not only provide global information to all input tokens but also group image tokens into categories. Based on category partitions, we further propose a category-based self-attention mechanism designed to leverage distant but similar tokens for enhancing input features. The experimental results show that our method achieves the best performance on various single image super-resolution benchmarks.

In this paper, we introduce a new perspective for improving image restoration by removing degradation in the textual representations of a given degraded image. Intuitively, restoration is much easier on text modality than image one. For example, it can be easily conducted by removing the degradation-related words while keeping the content-aware words. Hence, we combine the advantages of images in detail description and ones of text in degradation removal to perform restoration. To address the cross-modal assistance, we propose to map the degraded images into textual representations for removing the degradations, and then convert the restored textual representations into a guidance image for assisting image restoration. In particular, We ingeniously embed an image-to-text mapper and text restoration module into CLIP-equipped text-to-image models to generate the guidance. Then, we adopt a simple coarse-to-fine approach to dynamically inject multi-scale information from guidance to image restoration networks. Extensive experiments are conducted on various image restoration tasks, including deblurring, dehazing, deraining, and denoising, and all-in-one restoration. The results showcase that our method outperforms state-of-the-art ones across all these tasks. The codes and pre-trained models will be publicly available.

As an important and practical way to obtain high dynamic range (HDR) video, HDR video reconstruction from sequences with alternating exposures is still less explored, mainly due to the lack of large-scale real-world datasets. Existing methods are mostly trained on synthetic datasets, which perform poorly in real scenes. In this work, to facilitate the development of real-world HDR video reconstruction, we present Real-HDRV, a large-scale real-world benchmark dataset for HDR video reconstruction, featuring various scenes, diverse motion patterns, and high-quality labels. Specifically, our dataset contains 500 LDRs-HDRs video pairs, comprising about 28,000 LDR frames and 4,000 HDR labels, covering daytime, nighttime, indoor, and outdoor scenes. To our best knowledge, our dataset is the largest real-world HDR video reconstruction dataset. Correspondingly, we propose an end-to-end network for HDR video reconstruction, where a novel two-stage strategy is designed to perform alignment sequentially. Specifically, the first stage performs global alignment with the adaptively estimated global offsets, reducing the difficulty of subsequent alignment. The second stage implicitly performs local alignment in a coarse-to-fine manner at the feature level using the adaptive separable convolution. Extensive experiments demonstrate that: (1) models trained on our dataset can achieve better performance on real scenes than those trained on synthetic datasets; (2) our method outperforms previous state-of-the-art methods. Our dataset is available at https://github.com/yungsyu99/Real-HDRV.

Recovering images distorted by atmospheric turbulence is a challenging inverse problem due to the stochastic nature of turbulence. Although numerous turbulence mitigation (TM) algorithms have been proposed, their efficiency and generalization to real-world dynamic scenarios remain severely limited. Building upon the intuitions of classical TM algorithms, we present the Deep Atmospheric TUrbulence Mitigation network (DATUM). DATUM aims to overcome major challenges when transitioning from classical to deep learning approaches. By carefully integrating the merits of classical multi-frame TM methods into a deep network structure, we demonstrate that DATUM can efficiently perform long-range temporal aggregation using a recurrent fashion, while deformable attention and temporal-channel attention seamlessly facilitate pixel registration and lucky imaging. With additional supervision, tilt and blur degradation can be jointly mitigated. These inductive biases empower DATUM to significantly outperform existing methods while delivering a tenfold increase in processing speed. A large-scale training dataset, ATSyn, is presented as a co-invention to enable the generalization to real turbulence.

Pre-trained models with large-scale training data, such as CLIP and Stable Diffusion, have demonstrated remarkable performance in various high-level computer vision tasks such as image understanding and generation from language descriptions. Yet, their potential for low-level tasks such as image restoration remains relatively unexplored. In this paper, we explore such models to enhance image restoration. As off-the-shelf features (OSF) from pre-trained models do not directly serve image restoration, we propose to learn an additional lightweight module called Pre-Train-Guided Refinement Module (PTG-RM) to refine restoration results of a target restoration network with OSF. PTG-RM consists of two components, Pre-Train-Guided Spatial-Varying Enhancement (PTG-SVE), and Pre-Train-Guided Channel-Spatial Attention (PTG-CSA). PTG-SVE enables optimal short- and long-range neural operations, while PTG-CSA enhances spatial-channel attention for restoration-related learning. Extensive experiments demonstrate that PTG-RM, with its compact size ($<$1M parameters), effectively enhances restoration performance of various models across different tasks, including low-light enhancement, deraining, deblurring, and denoising.

This paper aims to address a common challenge in deep learning-based image transformation methods, such as image enhancement and super-resolution, which heavily rely on precisely aligned paired datasets with pixel-level alignments. However, creating precisely aligned paired images presents significant challenges and hinders the advancement of methods trained on such data. To overcome this challenge, this paper introduces a novel and simple Frequency Distribution Loss (FDL) for computing distribution distance within the frequency domain. Specifically, we transform image features into the frequency domain using Discrete Fourier Transformation (DFT). Subsequently, frequency components (amplitude and phase) are processed separately to form the FDL loss function. Our method is empirically proven effective as a training constraint due to the thoughtful utilization of global information in the frequency domain. Extensive experimental evaluations, focusing on image enhancement and super-resolution tasks, demonstrate that FDL outperforms existing misalignment-robust loss functions. Furthermore, we explore the potential of our FDL for image style transfer that relies solely on completely misaligned data. Our code is available at: https://github.com/eezkni/FDL

The performance of image restoration (IR) is highly dependent on the reconstruction quality of diverse contents with varying complexity. However, most IR approaches model the mapping between various complexity contents of inputs and outputs through the repeated feature calculation propagation mechanism in a unified pipeline, which leads to unsatisfactory results. To address this issue, we propose an explicit $\textbf{Co}$ntent $\textbf{De}$coupling framework for IR, dubbed $\textbf{CoDe}$, to end-to-end model the restoration process by utilizing decoupled content components in a divide-and-conquer-like architecture. Specifically, a Content Decoupling Module is first designed to decouple content components of inputs and outputs according to the frequency spectra adaptively generated from the transform domain. In addition, in order to harness the divide-and-conquer strategy for reconstructing decoupled content components, we propose an IR Network Container. It contains an optimized version, which is a streamlining of an arbitrary IR network, comprising the cascaded modulated subnets and a Reconstruction Layers Pool. Finally, a Content Consistency Loss is designed from the transform domain perspective to supervise the restoration process of each content component and further guide the feature fusion process. Extensive experiments on several IR tasks, such as image super-resolution, image denoising, and image blurring, covering both real and synthetic settings, demonstrate that the proposed paradigm can effectively take the performance of the original network to a new state-of-the-art level in multiple benchmark datasets (\eg, $\textbf{0.34}$dB@Set5 $\times4$ over DAT).

Generic Face Image Quality Assessment (GFIQA) evaluates the perceptual quality of facial images, which is crucial in improving image restoration algorithms and selecting high-quality face images for downstream tasks.We present a novel transformer-based method for GFIQA, which is aided by two unique mechanisms. First, a novel ``Dual-Set Degradation Representation Learning'' (DSL) mechanism uses facial images with both synthetic and real degradations to decouple degradation from content, ensuring generalizability to real-world scenarios. This self-supervised method learns degradation features on a global scale, providing a robust alternative to conventional methods that use local patch information in degradation learning. Second, our transformer leverages facial landmarks to emphasize visually salient parts of a face image in evaluating its perceptual quality.We also introduce a balanced and diverse Comprehensive Generic Face IQA (CGFIQA-40k) dataset of 40K images carefully designed to overcome the biases, in particular the imbalances in skin tone and gender representation, in existing datasets.Extensive analysis and evaluation demonstrate the robustness of our method, marking a significant improvement over prior methods. We will publish our code and dataset.

Recent image tone adjustment (or enhancement) approaches have predominantly adopted supervised learning for learning human-centric perceptual assessment. However, these approaches are constrained by intrinsic challenges of supervised learning. Primarily, the requirement for expertly-curated or retouched images escalates the data acquisition expenses. Moreover, their coverage of target style is confined to stylistic variants inferred from the training data. To surmount the above challenges, we propose an unsupervised learning-based approach for text-based image tone adjustment method, CLIPtone, that extends an existing image enhancement method to accommodate natural language descriptions. Specifically, we design a hypernetwork to adaptively modulate the pretrained parameters of the backbone model based on text description. To assess whether the adjusted image aligns with the text description without ground truth image, we utilize CLIP, which is trained on a vast set of language-image pairs and thus encompasses knowledge of human perception. Our approach offers numerous benefits such as a range of adjustments, minimal data collection expenses, and the ability for zero-shot predictions. While our work may bear similarities to existing text-based image editing and colorization methods, it stands out by preserving the contents of the original image, remaining lightweight and efficient in the adjustment process. Our approach’s efficacy is substantiated through comprehensive experiments, including a user study.

Transformer-based approaches have achieved promising performance in image restoration tasks, given their ability to model long-range dependencies, which is crucial for recovering clear images. Though diverse efficient attention mechanism designs have addressed the intensive computations associated with using transformers, they often involve redundant information and noisy interactions from irrelevant regions by considering all available tokens. In this work, we propose an Adaptive Sparse Transformer (AST) to mitigate the noisy interactions of irrelevant areas and remove feature redundancy in both spatial and channel domains. AST comprises two core designs, i.e., an Adaptive Sparse Self-Attention (ASSA) block and a Feature Refinement Feed-forward Network (FRFN). Specifically, ASSA is adaptively computed using a two-branch paradigm, where the sparse branch is introduced to filter out the negative impacts of low query-key matching scores for aggregating features, while the dense one ensures sufficient information flow through the network for learning discriminative representations. Meanwhile, FRFN employs an enhance-and-ease scheme to eliminate feature redundancy in channels, enhancing the restoration of clear latent images. Experimental results on commonly used benchmarks have demonstrated the versatility and competitive performance of our method in 6 tasks, including deraining, dehazing, deraindrop, demoireing, desnowing and deshadowing. The source code is provided in the supplementary materials.

Recently, numerous approaches have achieved notable success in compressed video quality enhancement (VQE). However, these methods usually ignore the utilization of valuable coding priors inherently embedded in compressed videos, such as motion vectors and residual frames, which carry abundant temporal and spatial information. To remedy this problem, we propose the Coding Priors-Guided Aggregation (CPGA) network to utilize temporal and spatial information from coding priors. The CPGA mainly consists of an inter-frame temporal aggregation (ITA) module and a multi-scale non-local aggregation (MNA) module. Specifically, the ITA module aggregates temporal information from consecutive frames and coding priors, while the MNA module globally captures spatial information guided by residual frames. In addition, to facilitate research in VQE task, we newly construct the Video Coding Priors (VCP) dataset, comprising 300 videos with various coding priors extracted from corresponding bitstreams. It remedies the shortage of previous datasets on the lack of coding information. Experimental results demonstrate the superiority of our method compared to existing state-of-the-art methods. The code and dataset will be released at https://github.com/VQE-CPGA/CPGA.

Adjusting camera exposure in arbitrary lighting conditions is the first step to ensure the functionality of computer vision applications.Poorly adjusted camera exposure often leads to critical failure and performance degradation.Traditional camera exposure control methods require multiple convergence steps and time-consuming processes, making them unsuitable for dynamic lighting conditions.In this paper, we propose a new camera exposure control framework that rapidly controls camera exposure while performing real-time processing by exploiting deep reinforcement learning.The proposed framework consists of four contributions: 1) a simplified training ground to simulate real-world's diverse and dynamic lighting changes, 2) flickering and image attribute-aware reward design, along with lightweight state design for real-time processing, 3) a static-to-dynamic lighting curriculum to gradually improve the agent's exposure-adjusting capability, and 4) domain randomization techniques to alleviate the limitation of the training ground and achieve seamless generalization in the wild. As a result, our proposed method rapidly reaches a desired exposure level within five steps with real-time processing (1 ms). Also, the acquired images are well-exposed and show superiority in various computer vision tasks, such as feature extraction and object detection.

Most of the previous exposure correction methods learn dense pixel-wise transformations to achieve promising results, but consume huge computational resources.Recently, Learnable 3D lookup tables (3D LUTs) have demonstrated impressive performance and efficiency for image enhancement.However, these methods can only perform global transformations and fail to finely manipulate local regions.Moreover, they uniformly downsample the input image, which loses the rich color information and limits the learning of color transformation capabilities.In this paper, we present a collaborative transformation framework (CoTF) for real-time exposure correction, which integrates global transformation with pixel-wise transformations in an efficient manner.Specifically, the global transformation adjusts the overall appearance using image-adaptive 3D LUTs to provide decent global contrast and sharp details, while the pixel transformation compensates for local context. Then, a relation-aware modulation module is designed to combine these two components effectively.In addition, we propose an adaptive sampling strategy to preserve more color information by predicting the sampling intervals, thus providing higher quality input data for the learning of 3D LUTs.Extensive experiments demonstrate that our method can process high-resolution images in real-time on GPUs while achieving comparable performance against current state-of-the-art methods.The code is available at https://github.com/HUST-IAL/CoTF.

Image warping, a classic task in computer vision, aims to use geometric transformations to change the appearance of images. Recent methods learn the resampling kernels for warping through neural networks to estimate missing values in irregular grids, which, however, fail to capture local variations in deformed content and produce images with distortion and less high-frequency details. To address this issue, this paper proposes an effective method, namely MFR, to learn Multi-Frequency Representations from input images for image warping. Specifically, we propose a progressive filtering network to learn image representations from different frequency subbands and generate deformable images in a coarse-to-fine manner. Furthermore, we employ learnable Gabor wavelet filters to improve the model's capability to learn local spatial-frequency representations. Comprehensive experiments, including homography transformation, equirectangular to perspective projection, and asymmetric image super-resolution, demonstrate that the proposed MFR significantly outperforms state-of-the-art image warping methods. Our method also showcases superior generalization to out-of-distribution domains, where the generated images are equipped with rich details and less distortion, thereby high visual quality.

Hyperspectral image (HSI) restoration aims at recovering clean images from degraded observations and plays a vital role in downstream tasks. Existing model-based methods have limitations in accurately modeling the complex image characteristics with handcraft priors, and deep learning-based methods suffer from poor generalization ability. To alleviate these issues, this paper proposes an unsupervised HSI restoration framework with pre-trained diffusion model (HIR-Diff), which restores the clean HSIs from the product of two low-rank components, i.e., the reduced image and the coefficient matrix. Specifically, the reduced image, which has a low spectral dimension, lies in the image field and can be inferred from our improved diffusion model where a new guidance function with total variation (TV) prior is designed to ensure that the reduced image can be well sampled. The coefficient matrix can be effectively pre-estimated based on singular value decomposition (SVD) and rank-revealing QR (RRQR) factorization. Furthermore, a novel exponential noise schedule is proposed to accelerate the restoration process (about 5$\times$ acceleration for denoising) with little performance decrease. Extensive experimental results validate the superiority of our method in both performance and speed on a variety of HSI restoration tasks, including HSI denoising, noisy HSI super-resolution, and noisy HSI inpainting.

This paper presents a novel zero-shot method for jointly denoising and enhancing real-word low-light images. The proposed method is independent of training data and noise distribution. Guided by illumination, we integrate denoising and enhancing processes seamlessly, enabling end-to-end training. Pairs of downsampled images are extracted from a single original low-light image and processed to preliminarily reduce noise. Based on the smoothness of illumination, near-authentic illumination can be estimated from the denoised low-light image. Specifically, the illumination is constrained by the denoised image's brightness, uniformly amplifying pixels to raise overall brightness to normal-light level. We simultaneously restrict the illumination by scaling each pixel of the denoised image based on its intensity, controlling the enhancement amplitude for different pixels. Applying the illumination to the original low-light image yields an adaptively enhanced reflection. This prevents under-enhancement and localized overexposure. Notably, we concatenate the reflection with the illumination, preserving their computational relationship, to ultimately remove noise from the original low-light image in the form of reflection. This provides sufficient image information for the denoising procedure without changing the noise characteristics. Extensive experiments demonstrate that our method outperforms other state-of-the-art methods. The source code is available at https://github.com/Doyle59217/ZeroIG.

We introduce a novel approach to single image denoising based on the Blind Spot Denoising principle, which we call MAsked and SHuffled Blind Spot Denoising (MASH). We focus on the case of correlated noise, which often plagues real images. MASH is the result of a careful analysis to determine the relationships between the level of blindness (masking) of the input and the (unknown) noise correlation. Moreover, we introduce a shuffling technique to weaken the local correlation of noise, which in turn yields an additional denoising performance improvement. We evaluate MASH via extensive experiments on real-world noisy image datasets. We demonstrate on par or better results compared to existing self-supervised denoising methods.

Most current arbitrary-scale image super-resolution (SR) methods has commonly relied on simulated data generated by simple synthetic degradation models (e.g., bicubic downsampling) at continuous various scales, thereby falling short in capturing the complex degradation of real-world images. This limitation hinders the visual quality of these methods when applied to real-world images. To address this issue, we propose the Continuous Optical Zooming dataset (COZ), by constructing an automatic imaging system to collect images at fine-grained various focal lengths within a specific range and providing strict image pair alignment. The COZ dataset serves as a benchmark to provide real-world data for training and testing arbitrary-scale SR models. To enhance the model's robustness against real-world image degradation, we propose a Local Mix Implicit network (LMI) based on the MLP-mixer architecture and meta-learning, which directly learns the local texture information by simultaneously mixing features and coordinates of multiple independent points. The extensive experiments demonstrate the superior performance of the arbitrary-scale SR models trained on the COZ dataset compared to models trained on simulated data. Our LMI model exhibits the superior effectiveness compared to other models. This study is of great significance in developing more efficient algorithms and improving the performance of arbitrary-scale image SR methods in practical applications. Our dataset and code will be publicly available.

While replacing Gaussian decoders with a conditional diffusion model enhances the perceptual quality of reconstructions in neural image compression, their lack of inductive bias for image data restricts their ability to achieve state-of-the-art perceptual levels. To address this limitation, we adopt a non-isotropic diffusion model at the decoder side. This model imposes an inductive bias aimed at distinguishing between frequency contents, thereby facilitating the generation of high-quality images. Moreover, our framework is equipped with a novel entropy model that accurately models the probability distribution of latent representation by exploiting spatio-channel correlations in latent space, while accelerating the entropy decoding step. This channel-wise entropy model leverages both local and global spatial contexts within each channel chunk. The global spatial context is built upon the Transformer, which is specifically designed for image compression tasks. The designed Transformer employs a Laplacian-shaped positional encoding, the learnable parameters of which are adaptively adjusted for each channel cluster. Our experiments demonstrate that our proposed framework yields better perceptual quality compared to cutting-edge generative-based codecs, and the proposed entropy model contributes to notable bitrate savings.

Event cameras offer many advantages over traditional frame-based cameras, such as high dynamic range and low latency. Therefore, event cameras are widely applied in diverse computer vision applications, where event-based keypoint detection is a fundamental task. However, achieving robust event-based keypoint detection remains challenging because the ground truth of event keypoints is difficult to obtain, descriptors extracted by CNN usually lack discriminative ability in the presence of intense noise, and fixed keypoint detectors are limited in detecting varied keypoint patterns. To address these challenges, a novel event-based keypoint detection method is proposed by learning dynamic detectors and contextual descriptors in a self-supervised manner (SD2Event), including a contextual feature descriptor learning (CFDL) module and a dynamic keypoint detector learning (DKDL) module. The proposed SD2Event enjoys several merits. First, the proposed CFDL module can model long-range contexts efficiently and effectively. Second, the DKDL module generates dynamic keypoint detectors, which can detect keypoints with diverse patterns across various event streams. Third, the proposed self-supervised signals can guide the model's adaptation to event data. Extensive experimental results on three challenging benchmarks show that our proposed method significantly outperforms state-of-the-art event-based keypoint detection methods.

This paper proposes LLaFS, the first attempt to leverage large language models (LLMs) in few-shot segmentation. In contrast to the conventional few-shot segmentation methods that only rely on the limited and biased information from the annotated support images, LLaFS leverages the vast prior knowledge gained by LLM as an effective supplement and directly uses the LLM to segment images in a few-shot manner. To enable the text-based LLM to handle image-related tasks, we carefully design an input instruction that allows the LLM to produce segmentation results represented as polygons, and propose a region-attribute table to simulate the human visual mechanism and provide multi-modal guidance. We also synthesize pseudo samples and use curriculum learning for pretraining to augment data and achieve better optimization. LLaFS achieves state-of-the-art results on multiple datasets, showing the potential of using LLMs for few-shot computer vision tasks.

While pre-trained large-scale vision models have shown significant promise for semantic correspondence, their features often struggle to grasp the geometry and orientation of instances.This paper identifies the importance of being geometry-aware for semantic correspondence and reveals a limitation of the features of current foundation models under simple post-processing.We show that incorporating this information can markedly enhance semantic correspondence performance with simple but effective solutions in both zero-shot and supervised settings. We also construct a new challenging benchmark for semantic correspondence built from an existing animal pose estimation dataset, for both pre-training validating models. Our method achieves a PCK@0.10 score of 65.4 (zero-shot) and 85.6 (supervised) on the challenging SPair-71k dataset, outperforming the state of the art by 5.5p and 11.0p absolute gains, respectively.Our code and datasets are publicly available at: https://telling-left-from-right.github.io.

We introduce One-shot Open Affordance Learning (OOAL), where a model is trained with just one example per base object category, but is expected to identify novel objects and affordances. While vision-language models excel at recognizing novel objects and scenes, they often struggle to understand finer levels of granularity such as affordances. To handle this issue, we conduct a comprehensive analysis of existing foundation models, to explore their inherent understanding of affordances and assess the potential for data-limited affordance learning. We then propose a vision-language framework with simple and effective designs that boost the alignment between visual features and affordance text embeddings. Experiments on two affordance segmentation benchmarks show that the proposed method outperforms state-of-the-art models with less than 1\% of the full training data, and exhibits reasonable generalization capability on unseen objects and affordances.

This paper presents a simple but performant semi-supervised semantic segmentation approach, called CorrMatch. Previous approaches mostly employ complicated training strategies to leverage unlabeled data but overlook the role of correlation maps in modeling the relationships between pairs of locations. We observe that the correlation maps not only enable clustering pixels of the same category easily but also contain good shape information, which previous works have omitted. Motivated by these, we aim to improve the use efficiency of unlabeled data by designing two novel label propagation strategies. First, we propose to conduct pixel propagation by modeling the pairwise similarities of pixels to spread the high-confidence pixels and dig out more. Then, we perform region propagation to enhance the pseudo labels with accurate class-agnostic masks extracted from the correlation maps. CorrMatch achieves great performance on popular segmentation benchmarks. Taking the DeepLabV3+ with ResNet-101 backbone as our segmentation model, we receive a 76\%+ mIoU score on the Pascal VOC 2012 dataset with only 92 annotated images. The source code will be made publicly available.

Domain Generalized Semantic Segmentation (DGSS) deals with training a model on a labeled source domain with the aim of generalizing to unseen domains during inference. Existing DGSS methods typically effectuate robust features by means of Domain Randomization (DR). Such an approach is often limited as it can only account for style diversification and not content. In this work, we take an orthogonal approach to DGSS and propose to use an assembly of CoLlaborative FOUndation models for Domain Generalized Semantic Segmentation (CLOUDS). In detail, CLOUDS is a framework that integrates Foundation Models of various kinds: (i) CLIP backbone for its robust feature representation, (ii) Diffusion Model to diversify the content, thereby covering various modes of the possible target distribution, and (iii) Segment Anything Model (SAM) for iteratively refining the predictions of the segmentation model. Extensive experiments show that our CLOUDS excels in adapting from synthetic to real DGSS benchmarks and under varying weather conditions, notably outperforming prior methods by 5.6% and 6.7% on averaged mIoU, respectively. Our code is available at https://github.com/yasserben/CLOUDS

The Segment Anything Model (SAM) marks a notable milestone in segmentation models, highlighted by its robust zero-shot capabilities and ability to handle diverse prompts. SAM follows a pipeline that separates interactive segmentation into image preprocessing through a large encoder and interactive inference via a lightweight decoder, ensuring efficient real-time performance. However, SAM faces stability issues in challenging samples upon this pipeline. These issues arise from two main factors. Firstly, the image preprocessing disables SAM to dynamically use image-level zoom-in strategies to refocus on the target object during interaction. Secondly, the lightweight decoder struggles to sufficiently integrate interactive information with image embeddings. To address these two limitations, we propose FocSAM with a pipeline redesigned on two pivotal aspects. First, we propose Dynamic Window Multi-head Self-Attention (Dwin-MSA) to dynamically refocus SAM's image embeddings on the target object. Dwin-MSA localizes attention computations around the target object, enhancing object-related embeddings with minimal computational overhead. Second, we propose Pixel-wise Dynamic ReLU (P-DyReLU) to enable sufficient integration of interactive information from a few initial clicks that have significant impacts on the overall segmentation results. Experimentally, FocSAM augments SAM’s interactive segmentation performance to match the existing state-of-the-art method in segmentation quality, requiring only about $5.6\%$ of this method's inference time on CPUs. Code is available at https://github.com/YouHuang67/focsam.

The Laplace-Beltrami operator (LBO) emerges from studying manifolds equipped with a Riemannian metric. It is often called the "swiss army knife of geometry processing" as it allows to capture intrinsic shape information and gives rise to heat diffusion, geodesic distances, and a multitude of shape descriptors. It also plays a central role in geometric deep learning. In this work, we explore Finsler manifolds as a generalization of Riemannian manifolds. We revisit the Finsler heat equation and derive a Finsler heat kernel and a Finsler-Laplace-Beltrami Operator (FLBO): a novel theoretically justified anisotropic Laplace-Beltrami operator (ALBO). In experimental evaluations we demonstrate that the proposed FLBO is a valuable alternative to the traditional Riemannian-based LBO and ALBOs for spatial filtering and shape correspondence estimation. We hope that the proposed Finsler heat kernel and the FLBO will inspire further exploration of Finsler geometry in the computer vision community.

We tackle the problem of building digital twins of unknown articulated objects from two RGBD scans of the object at different articulation states. We decompose the problem into two stages, each addressing distinct aspects. Our method first reconstructs object-level shape at each state, then recovers the underlying articulation model including part segmentation and joint articulations that associate the two states. By explicitly modeling point-level correspondences and exploiting cues from images, 3D reconstructions, and kinematics, our method yields more accurate and stable results compared to prior work. It also handles more than one movable part and does not rely on any object shape or structure priors. Project page: https://digitaltwinart.github.io/digitaltwinart2023/

We present Cutie, a video object segmentation (VOS) network with object-level memory reading, which puts the object representation from memory back into the video object segmentation result. Recent works on VOS employ bottom-up pixel-level memory reading which struggles due to matching noise, especially in the presence of distractors, resulting in lower performance in more challenging data. In contrast, Cutie performs top-down object-level memory reading by adapting a small set of object queries. Via those, it interacts with the bottom-up pixel features iteratively with a query-based object transformer (qt, hence Cutie). The object queries act as a high-level summary of the target object, while high-resolution feature maps are retained for accurate segmentation. Together with foreground-background masked attention, Cutie cleanly separates the semantics of the foreground object from the background. On the challenging MOSE dataset, Cutie improves by 8.7 J&F over XMem with a similar running time and improves by 4.2 J&F over DeAOT while being three times faster. Code will be released.

In this paper, we address the challenge of image resolution variation for the Segment Anything Model (SAM). SAM, known for its zero-shot generalizability, exhibits a performance degradation when faced with datasets with varying image sizes. Previous approaches tend to resize the image to a fixed size or adopt structure modifications, hindering the preservation of SAM's rich prior knowledge. Besides, such task-specific tuning necessitates a complete retraining of the model, which is cost-expensive and unacceptable for deployment in the downstream tasks. In this paper, we reformulate this issue as a length extrapolation problem, where token sequence length varies while maintaining a consistent patch size for images of different sizes. To this end, we propose a Scalable Bias-Mode Attention Mask (BA-SAM) to enhance SAM's adaptability to varying image resolutions while eliminating the need for structure modifications. Firstly, we introduce a new scaling factor to ensure consistent magnitude in the attention layer's dot product values when the token sequence length changes. Secondly, we present a bias-mode attention mask that allows each token to prioritize neighboring information, mitigating the impact of untrained distant information. Our BA-SAM demonstrates efficacy in two scenarios: zero-shot and fine-tuning. Extensive evaluation on diverse datasets, including DIS5K, DUTS, ISIC, COD10K, and COCO, reveals its ability to significantly mitigate performance degradation in the zero-shot setting and achieve state-of-the-art performance with minimal fine-tuning. Furthermore, we propose a generalized model and benchmark, showcasing BA-SAM's generalizability across all four datasets simultaneously.

Part-aware panoptic segmentation (PPS) requires (a) that each foreground object and background region in an image is segmented and classified, and (b) that all parts within foreground objects are segmented, classified and linked to their parent object. Existing methods approach PPS by separately conducting object-level and part-level segmentation. However, their part-level predictions are not linked to individual parent objects. Therefore, their learning objective is not aligned with the PPS task objective, which harms the PPS performance. To solve this, and make more accurate PPS predictions, we propose Task-Aligned Part-aware Panoptic Segmentation (TAPPS). This method uses a set of shared queries to jointly predict (a) object-level segments, and (b) the part-level segments within those same objects. As a result, TAPPS learns to predict part-level segments that are linked to individual parent objects, aligning the learning objective with the task objective, and allowing TAPPS to leverage joint object-part representations. With experiments, we show that TAPPS considerably outperforms methods that predict objects and parts separately, and achieves new state-of-the-art PPS results.

Interpreting camera data is key for autonomously-acting systems, such as autonomous vehicles. Vision systems that operate in real-world environments must be able to understand their surroundings and need the ability to deal with novel situations. This paper tackles open-world semantic segmentation, i.e., the variant of interpreting camera data in which objects occur that have not been seen during training. We propose a novel approach that performs accurate closed-world semantic segmentation and, at the same time, can identify new categories without requiring any additional training data. Our approach additionally provides a similarity measure for every newly discovered class in an image to a known category, which is useful information in further downstream tasks such as planning or mapping. Through extensive experiments, we show that our model achieves state-of-the-art results on classes known from training data as well as for anomaly segmentation and can distinguish between different unknown classes.

Superpixels play a crucial role in image processing by partitioning an image into clusters of pixels with similar visual attributes.This facilitates subsequent image processing tasks, offering computational advantages over the manipulation of individual pixels. While numerous oversegmentation techniques have emerged in recent years, many rely on predefined initialization and termination criteria.In this paper, a novel top-down superpixel segmentation algorithm called Hierarchical Histogram Threshold Segmentation (HHTS) is introduced.It eliminates the need for initialization and implements auto-termination, outperforming state-of-the-art methods w.r.t boundary recall.This is achieved by iteratively partitioning individual pixel segments into foreground and background and applying intensity thresholding across multiple color channels.The underlying iterative process constructs a superpixel hierarchy that adapts to local detail distributions until color information exhaustion.Experimental results demonstrate the superiority of the proposed approach in terms of boundary adherence, while maintaining competitive runtime performance on the BSDS500 and NYUV2 datasets.Furthermore, an application of HHTS in refining machine learning-based semantic segmentation masks produced by the Segment Anything Foundation Model (SAM) is presented.

Powered by massive curated training data, Segment Anything Model (SAM) has demonstrated its impressive generalization capabilities in open-world scenarios with the guidance of manual prompts. However, the vanilla SAM is class-agnostic and heavily relies on user-provided prompts to segment objects of interest. Customizing it into diversified tasks becomes necessary to identify specific targets, preventing suboptimal segmentation performance. In this paper, we propose a novel framework, termed AlignSAM, designed for automatic prompting to align SAM within an open context through reinforcement learning. Anchored by an agent, AlignSAM enables the generality of the SAM model across diverse downstream tasks while keeping its parameters frozen. Specifically, AlignSAM initiates a prompting agent to iteratively refine segmentation predictions by interacting with the foundational model. It integrates an additional reinforcement learning network to provide informative prompts to the foundational models. Additionally, a semantic recalibration module is introduced to provide prompt labels, enhancing the agent's proficiency in handling tasks encompassing explicit and implicit semantics. Experiments conducted on various challenging segmentation tasks among existing foundation models demonstrate the superiority of the proposed AlignSAM over state-of-the-art approaches.

Recently, the Segment Anything Model (SAM) has showcased remarkable capabilities of zero-shot segmentation, while NeRF (Neural Radiance Fields) has gained popularity as a method for various 3D problems beyond novel view synthesis. Though there exist initial attempts to incorporate these two methods into 3D segmentation, they face the challenge of accurately and consistently segmenting objects in complex scenarios. In this paper, we introduce the Segment Anything for NeRF in High Quality (SANeRF-HQ) to achieve high-quality 3D segmentation of any target object in a given scene. SANeRF-HQ utilizes SAM for open-world object segmentation guided by user-supplied prompts, while leveraging NeRF to aggregate information from different viewpoints. To overcome the aforementioned challenges, we employ density field and RGB similarity to enhance the accuracy of segmentation boundary during the aggregation. Emphasizing on segmentation accuracy, we evaluate our method on multiple NeRF datasets where high-quality ground-truths are available or manually annotated. SANeRF-HQ shows a significant quality improvement over state-of-the-art methods in NeRF object segmentation, provides higher flexibility for object localization, and enables more consistent object segmentation across multiple views. Results and code are available at the project site: https://lyclyc52.github.io/SANeRF-HQ/

Despite the recent advances in unified image segmentation (IS), developing a unified video segmentation (VS) model remains a challenge. This is mainly because generic category-specified VS tasks need to detect all objects and track them across consecutive frames, while prompt-guided VS tasks require re-identifying the target with visual/text prompts throughout the entire video, making it hard to handle the different tasks with the same architecture. We make an attempt to address these issues and present a novel unified VS architecture, namely UniVS, by using prompts as queries. UniVS averages the prompt features of the target from previous frames as its initial query to explicitly decode masks, and introduces a target-wise prompt cross-attention layer in the mask decoder to integrate prompt features in the memory pool. By taking the predicted masks of entities from previous frames as their visual prompts, UniVS converts different VS tasks into prompt-guided target segmentation, eliminating the heuristic inter-frame matching process. Our framework not only unifies the different VS tasks but also naturally achieves universal training and testing, ensuring robust performance across different scenarios. UniVS shows a commendable balance between performance and universality on 10 challenging VS benchmarks, covering video instance, semantic, panoptic, object, and referring segmentation tasks. Code can be found at \url{https://github.com/MinghanLi/UniVS}.

Detecting edges in images suffers from the problems of (P1) heavy imbalance between positive and negative classes as well as (P2) label uncertainty owing to disagreement between different annotators. Existing solutions address P1 using class-balanced cross-entropy loss and dice loss and P2 by only predicting edges agreed upon by most annotators. In this paper, we propose RankED, a unified ranking-based approach that addresses both the imbalance problem (P1) and the uncertainty problem (P2). RankED tackles these two problems with two components: One component which ranks positive pixels over negative pixels, and the second which promotes high confidence edge pixels to have more label certainty. We show that RankED outperforms previous studies and sets a new state-of-the-art on NYUD-v2, BSDS500 and Multi-cue datasets. Code is available at https://ranked-cvpr24.github.io.

In the field of video object segmentation (VOS), low-light conditions pose a significant challenge, often leading to extremely poor image quality and inaccurate matching by computing similarity between query and memory frames. Event cameras have high dynamic range and motion information of objects. These characteristics provide visibility of objects to assist VOS methods under low-light conditions. In this paper, we introduce a novel framework for low-light VOS, incorporating event camera data to improve segmentation accuracy. Our approach consists of two key components: Event-Guided Memory Matching (EGMM) and Adaptive Cross-Modal Fusion (ACMF). The EGMM module is designed to solve the inaccurate matching under low-light conditions. On the other hand, the ACMF module aims to extract valuable features from noise features by adaptive fusing image and event modalities. Besides, we construct a simulated Low-Light Event DAVIS (LLE-DAVIS) dataset and collect a real-world Low-Light Event Object Segmentation (LL-EOS) dataset, including frames and events. Experiments validate the effectiveness of our method on both datasets.

Densely annotating the large-scale point clouds is laborious. To alleviate the annotation burden, contrastive learning has attracted increasing attention for tackling semi-supervised 3D semantic segmentation. However, existing point-to-point contrastive learning techniques in literature are generally sensitive to outliers, resulting in insufficient modeling of the point-wise representations. To address this problem, we propose a method named DDSemi for semi-supervised 3D semantic segmentation, where a density-guided contrastive learning technique is explored. This technique calculates the contrastive loss in a point-to-anchor manner by estimating an anchor for each class from the memory bank based on the finding that the cluster centers tend to be located in dense regions. In this technique, an inter-contrast loss is derived from the perturbed unlabeled point cloud pairs, while an intra-contrast loss is derived from a single unlabeled point cloud. The derived losses could enhance the discriminability of the features and implicitly constrain the semantic consistency between the perturbed unlabeled point cloud pairs. In addition, we propose a dual-space hardness sampling strategy to pay more attention to the hard samples located in sparse regions of both the geometric space and feature space by reweighting the point-wise intra-contrast loss. Experimental results on both indoor-scene and outdoor-scene datasets demonstrate that the proposed method outperforms the comparative state-of-the-art semi-supervised methods.

CLIP has demonstrated marked progress in visual recognition due to its powerful pre-training on large-scale image-text pairs. However, it still remains a critical challenge: how to transfer image-level knowledge into pixel-level understanding tasks such as semantic segmentation. In this paper, to solve the mentioned challenge, we analyze the gap between the capability of the CLIP model and the requirement of the zero-shot semantic segmentation task. Based on our analysis and observations, we propose a novel method for zero-shot semantic segmentation, dubbed CLIP-RC (CLIP with Regional Clues), bringing two main insights. On the one hand, a region-level bridge is necessary to provide fine-grained semantics. On the other hand, overfitting should be mitigated during the training stage. Benefiting from the above discoveries, CLIP-RC achieves state-of-the-art performance on various zero-shot semantic segmentation benchmarks, including PASCAL VOC, PASCAL Context, and COCO-Stuff 164K. Code will be available at https://github.com/Jittor/JSeg.

Shape assembly composes complex shapes geometries by arranging simple part geometries and has wide applications in autonomous robotic assembly and CAD modeling. Existing works focus on geometry reasoning and neglect the actual physical assembly process of matching and fitting joints, which are the contact surfaces connecting different parts. In this paper, we consider contacting joints for the task of multi-part assembly. A successful joint-optimized assembly needs to satisfy the bilateral objectives of shape structure and joint alignment. We propose a hierarchical graph learning approach composed of two levels of graph representation learning. The part graph takes part geometries as input to build the desired shape structure. The joint-level graph uses part joints information and focuses on matching and aligning joints. The two kinds of information are combined to achieve the bilateral objectives. Extensive experiments demonstrate that our method outperforms previous methods, achieving better shape structure and higher joint alignment accuracy.

Advancements in 3D instance segmentation have traditionally been tethered to the availability of annotated datasets, limiting their application to a narrow spectrum of object categories. Recent efforts have sought to harness vision-language models like CLIP for open-set semantic reasoning, yet these methods struggle to distinguish between objects of the same categories and rely on specific prompts that are not universally applicable. In this paper, we introduce SAI3D, a novel zero-shot 3D instance segmentation approach that synergistically leverages geometric priors and semantic cues derived from Segment Anything Model (SAM). Our method partitions a 3D scene into geometric primitives, which are then progressively merged into 3D instance segmentations that are consistent with the multi-view SAM masks. Moreover, we design a hierarchical region-growing algorithm with a dynamic thresholding mechanism, which largely improves the robustness of fine-grained 3D scene parsing. Empirical evaluations on ScanNet, Matterport3D and the more challenging ScanNet++ datasets demonstrate the superiority of our approach. Notably, SAI3D outperforms existing open-vocabulary baselines and even surpasses fully-supervised methods in class-agnostic segmentation on ScanNet++. Our project page is at https://yd-yin.github.io/SAI3D/.

Semi-supervised semantic segmentation allows model to mine effective supervision from unlabeled data to complement label-guided training. Recent research has primarily focused on consistency regularization techniques, exploring perturbation-invariant training at both the image and feature levels. In this work, we proposed a novel feature-level consistency learning framework named Density-Descending Feature Perturbation (DDFP). Inspired by the low-density separation assumption in semi-supervised learning, our key insight is that feature density can shed a light on the most promising direction for the segmentation classifier to explore, which is the regions with lower density. We propose to shift features with confident predictions towards lower-density regions by perturbation injection. The perturbed features are then supervised by the predictions on the original features, thereby compelling the classifier to explore less dense regions to effectively regularize the decision boundary. Central to our method is the estimation of feature density. To this end, we introduce a lightweight density estimator based on normalizing flow, allowing for efficient capture of the feature density distribution in an online manner. By extracting gradients from the density estimator, we can determine the direction towards less dense regions for each feature. The proposed DDFP outperforms other designs on feature-level perturbations and shows state of the art performances on both Pascal VOC and Cityscapes dataset under various partition protocols. The project is available at https://github.com/Gavinwxy/DDFP.

Non-isometric shape correspondence remains a fundamental challenge in computer vision. Traditional methods using Laplace-Beltrami operator (LBO) eigenmodes face limitations in characterizing high-frequency extrinsic shape changes like bending and creases. We propose a novel approach of combining the non-orthogonal extrinsic basis of eigenfunctions of the elastic thin-shell hessian with the intrinsic ones of the LBO, creating a hybrid spectral space in which we construct functional maps. To this end, we present a theoretical framework to effectively integrate non-orthogonal basis functions into descriptor- and learning-based functional map methods. Our approach can be incorporated easily into existing functional map pipelines across varying applications and is able to handle complex deformations beyond isometries. We show extensive evaluations across various supervised and unsupervised settings and demonstrate significant improvements. Notably, our approach achieves up to 15\% better mean geodesic error for non-isometric correspondence settings and up to 45\% improvement in scenarios with topological noise. Code will be made available upon acceptance.

Recent weakly supervised semantic segmentation (WSSS) methods strive to incorporate contextual knowledge to improve the completeness of class activation maps (CAM). In this work, we argue that the knowledge bias between instances and contexts affects the capability of the prototype to sufficiently understand instance semantics. Inspired by prototype learning theory, we propose leveraging prototype awareness to capture diverse and fine-grained feature attributes of instances. The hypothesis is that contextual prototypes might erroneously activate similar and frequently co-occurring object categories due to this knowledge bias. Therefore, we propose to enhance the prototype representation ability by mitigating the bias to better capture spatial coverage in semantic object regions. With this goal, we present a Context Prototype-Aware Learning (CPAL) strategy, which leverages semantic context to enrich instance comprehension. The core of this method is to accurately capture intra-class variations in object features through context-aware prototypes, facilitating the adaptation to the semantic attributes of various instances. We design feature distribution alignment to optimize prototype awareness, aligning instance feature distributions with dense features. In addition, a unified training framework is proposed to combine label-guided classification supervision and prototypes-guided self-supervision. Experimental results on PASCAL VOC 2012 and MS COCO 2014 show that CPAL significantly improves off-the-shelf methods and achieves state-of-the-art performance.

Model calibration, measuring the alignment between the prediction accuracy and model confidence, is an important metric reflecting model trustworthiness. Existing dense binary classification methods, without proper regularisation of model confidence, are prone to being over-confident. To calibrate Deep Neural Networks (DNNs), we propose a Self-Calibrating Vicinal Risk Minimisation (SCVRM) that explores the vicinity space of labeled data, where vicinal images that are farther away from labeled images adopt the groundtruth label with decreasing label confidence. We prove that in the logistic regression problem, SCVRM can be seen as a Vicinal Risk Minimisation plus a regularisation term that penalises the over-confident predictions. In practical implementation, SCVRM is approximated with Monte Carlo sampling that samples additional augmented training images from the vicinal distributions. Experimental results demonstrate that SCVRM can significantly enhance model calibration for different dense classification tasks on both in-distribution and out-of-distribution data. Code will be released.

In the field of computer vision, panoptic segmentation, which merges semantic and instance segmentation, stands as a cutting-edge task. Despite the recent significant progress in panoptic segmentation with deep learning models, the dynamic nature of real-world applications necessitates continual learning, where models adapt to new classes (plasticity) over time without forgetting old ones (catastrophic forgetting). Current continual segmentation methods often rely on distillation strategies like knowledge distillation and pseudo-labeling, which are effective but result in increased training complexity and computational overhead. In this paper, we introduce a novel and efficient method for continual panoptic segmentation based on Visual Prompt Tuning (VPT), dubbed ECLIPSE. Our approach involves freezing the base model parameters and fine-tuning only a small set of prompt embeddings, addressing both catastrophic forgetting and plasticity and significantly reducing the trainable parameters. To mitigate inherent challenges such as error propagation and semantic drift in continual segmentation, we propose logit manipulation to effectively leverage common knowledge across the classes.Experiments on ADE20K continual panoptic segmentation benchmark demonstrate the superiority of ECLIPSE, notably its robustness against catastrophic forgetting and its reasonable plasticity, achieving a new state-of-the-art.

Prominent solutions for medical image segmentation are typically tailored for automatic or interactive setups, posing challenges in facilitating progress achieved in one task to another. This also necessitates separate models for each task, duplicating both training time and parameters. To address above issues, we introduce S2VNet, a universal framework that leverages Slice-to-Volume propagation to unify automatic/interactive segmentation within a single model and one training session. Inspired by clustering-based segmentation techniques, S2VNet makes full use of the slice-wise structure of volumetric data by initializing cluster centers from the cluster results of previous slice. This enables knowledge acquired from prior slices to assist in the segmentation of the current slice, further efficiently bridging the communication between remote slices using mere 2D networks. Moreover, such a framework readily accommodates interactive segmentation with no architectural change, simply by initializing centroids from user inputs. S2VNet distinguishes itself by swift inference speeds and reduced memory consumption compared to prevailing 3D solutions. It can also handle multi-class interactions with each of them serving to initialize different centroids. Experiments on three benchmarks demonstrate S2VNet surpasses task-specified solutions on both automatic/interactive setups.

Existing few-shot segmentation methods usually extract foreground prototypes from support images to guide query image segmentation. However, different background contexts of support and query images can cause their foreground features to be misaligned. This phenomenon, known as background context bias, can hinder the effectiveness of support prototypes in guiding query image segmentation. In this work, we propose a novel framework with an iterative structure to address this problem. In each iteration of the framework, we first generate a query prediction based on a support foreground feature. Next, we extract background context from the query image to modulate the support foreground feature, thus eliminating the foreground feature misalignment caused by the different backgrounds. After that, we design a confidence-biased attention to eliminate noise and cleanse information. By integrating these components through an iterative structure, we create a novel network that can leverage the synergies between different modules to improve their performance in a mutually reinforcing manner. Through these carefully designed components and structures, our network can effectively eliminate background context bias in few-shot segmentation, thus achieving outstanding performance. We conduct extensive experiments on the PASCAL-$5^{i}$ and COCO-$20^{i}$ datasets and achieve state-of-the-art (SOTA) results, which demonstrate the effectiveness of our approach.

Cross-Domain Few-Shot Segmentation (CD-FSS) poses the challenge of segmenting novel categories from a distinct domain using only limited exemplars. In this paper, we undertake a comprehensive study of CD-FSS and uncover two crucial insights: (i) the necessity of a fine-tuning stage to effectively transfer the learned meta-knowledge across domains, and (ii) the overfitting risk during the naïve fine-tuning due to the scarcity of novel category examples. With these insights, we propose a novel cross-domain fine-tuning strategy that addresses the challenging CD-FSS tasks. We first design Bi-directional Few-shot Prediction (BFP), which establishes support-query correspondence in a bi-directional manner, crafting augmented supervision to reduce the overfitting risk. Then we further extend BFP into Iterative Few-shot Adaptor (IFA), which is a recursive framework to capture the support-query correspondence iteratively, targeting maximal exploitation of supervisory signals from the sparse novel category samples. Extensive empirical evaluations show that our method significantly outperforms the state-of-the-arts (+7.8%), which verifies that IFA tackles the cross-domain challenges and mitigates the overfitting simultaneously. The code is available at: https://github.com/niejiahao1998/IFA.

The key lie in semi-supervised semantic segmentation is how to fully exploit substantial unlabeled data to improve the model’s generalization performance by resorting to constructing effective supervision signals. Most methods tend to directly apply contrastive learning to seek additional supervision to complement independent regular pixel-wise consistency regularization. However, these methods tend not to be preferred ascribed to their complicated designs, heavy memory footprints and susceptibility to confirmation bias. In this paper, we analyze the bottlenecks exist in contrastive learning-based methods and offer a fresh perspective on inter-pixel correlations to construct more safe and effective supervision signals, which is in line with the nature of semantic segmentation. To this end, we develop a coherent RankMatch network, including the construction of representative agents to model inter-pixel correlation beyond regular individual pixel-wise consistency, and further unlock the potential of agents by modeling inter-agent relationships in pursuit of rank-aware correlation consistency. Extensive experimental results on multiple benchmarks, including mitochondria segmentation, demonstrate that RankMatch performs favorably against state-of-the-art methods. Particularly in the low-data regimes, RankMatch achieves significant improvements.

Audiovisual segmentation (AVS) is a challenging task that aims to segment visual objects in videos according to their associated acoustic cues. With multiple sound sources and background disturbances involved, establishing robust correspondences between audio and visual contents poses unique challenges due to its (1) complex entanglement across sound sources and (2) frequent changes in the occurrence of distinct sound events. Assuming sound events occur independently, the multi-source semantic space can be represented as the Cartesian product of single-source sub-spaces.We are motivated to decompose the multi-source audio semantics into single-source semantics for more effective interactions with visual content.We propose a semantic decomposition method based on product quantization, where the multi-source semantics can be decomposed and represented by several disentangled and noise-suppressed single-source semantics. Furthermore, we introduce a global-to-local quantization mechanism, which distills knowledge from stable global (clip-level) features into local (frame-level) ones, to handle frequent changes in audio semantics.Extensive experiments demonstrate that our semantically decomposed audio representation significantly improves AVS performance, e.g., +21.2% mIoU on the challenging AVS-Semantic benchmark with ResNet50 backbone.

Dilated convolution, which expands the receptive field by inserting gaps between its consecutive elements, is widely employed in computer vision. In this study, we propose three strategies to improve individual phases of dilated convolution from the view of spectrum analysis. Departing from the conventional practice of fixing a global dilation rate as a hyperparameter, we introduce Frequency-Adaptive Dilated Convolution (FADC) which dynamically adjusts dilation rates spatially based on local frequency components. Subsequently, we design two plug-in modules to directly enhance effective bandwidth and receptive field size.The Adaptive Kernel (AdaKern) module decomposes convolution weights into low-frequency and high-frequency components, dynamically adjusting the ratio between these components on a per-channel basis. By increasing the high-frequency part of convolution weights, AdaKern captures more high-frequency components, thereby improving effective bandwidth.The Frequency Selection (FreqSelect) module optimally balances high- and low-frequency components in feature representations through spatially variant reweighting. It suppresses high frequencies in the background to encourage FADC to learn a larger dilation, thereby increasing the receptive field for an expanded scope. Extensive experiments on segmentation and object detection consistently validate the efficacy of our approach. The code is made publicly available at https://github.com/Linwei-Chen/FADC.

Open-vocabulary semantic segmentation strives to distinguish pixels into different semantic groups from an open set of categories. Most existing methods explore utilizing pre-trained vision-language models, in which the key is to adapt the image-level model for pixel-level segmentation task. In this paper, we propose a simple encoder-decoder, named SED, for open-vocabulary semantic segmentation, which comprises a hierarchical encoder-based cost map generation and a gradual fusion decoder with category early rejection. The hierarchical encoder-based cost map generation employs hierarchical backbone, instead of plain transformer, to predict pixel-level image-text cost map. Compared to plain transformer, hierarchical backbone better captures local spatial information and has linear computational complexity with respect to input size. Our gradual fusion decoder employs a top-down structure to combine cost map and the feature maps of different backbone levels for segmentation. To accelerate inference speed, we introduce a category early rejection scheme in the decoder that rejects many no-existing categories at the early layer of decoder, resulting in at most 4.7 times acceleration without accuracy degradation. Experiments are performed on multiple open-vocabulary semantic segmentation datasets, which demonstrates the efficacy of our SED method. When using ConvNeXt-B, our SED method achieves mIoU score of 31.6\% on ADE20K with 150 categories at 82 millisecond ($ms$) per image on a single A6000. Our source code is available at https://github.com/xb534/SED.

Image-level Weakly Supervised Semantic Segmentation (WSSS) has received increasing attention due to its low annotation cost. Class Activation Mapping (CAM) generated through classifier weights in WSSS inevitably ignores certain useful cues, while the CAM generated through class prototypes can alleviate that. However, because of the different goals of image classification and semantic segmentation, the class prototypes still focus on activating primary discriminative pixels learned from classification loss, leading to incomplete CAM. In this paper, we propose a plug-and-play Prototype-based Secondary Discriminative Pixels Mining (PSDPM) framework for enabling class prototypes to activate more secondary discriminative pixels, thus generating a more complete CAM. Specifically, we introduce a Foreground Pixel Estimation Module (FPEM) for estimating potential foreground pixels based on the correlations between primary and secondary discriminative pixels and the semantic segmentation results of baseline methods. Then, we enable WSSS model to learn discriminative features from secondary discriminative pixels through a consistency loss calculated between FPEM result and class-prototype CAM. Experimental results show that our PSDPM improves various baseline methods significantly and achieves new state-of-the-art performances on WSSS benchmarks. Codes are available at https://github.com/xinqiaozhao/PSDPM.

Point cloud matching, a crucial technique in computer vision, medical and robotics fields, is primarily concerned with finding correspondences between pairs of point clouds or voxels. In some practical scenarios, emphasizing local differences is crucial for accurately identifying a correct match, thereby enhancing the overall robustness and reliability of the matching process. Commonly used shape descriptors have several limitations and often fail to provide meaningful local insights about the paired geometries. In this work, we propose a new technique, based on graph Laplacian eigenmaps, to match point clouds by taking into account fine local structures. To deal with the order and sign ambiguity of Laplacian eigenmaps, we introduce a new operator, called Coupled Laplacian, that allows to easily generate aligned eigenspaces for multiple registered geometries. We show that the similarity between those aligned high-dimensional spaces provides a locally meaningful score to match shapes. We firstly evaluate the performance of the proposed technique in a point-wise manner, focusing on the task of object anomaly localization on the MVTec 3D-AD dataset. Additionally, we define a new medical task, called automatic Bone Side Estimation (BSE), which we address through a global similarity score derived from coupled eigenspaces. In order to test it, we propose a benchmark collecting bone surface structures from various public datasets. Our matching technique, based on Coupled Laplacian, outperforms other methods by reaching an impressive accuracy on both tasks.

This paper aims to achieve universal segmentation of arbitrary semantic level.Despite significant progress in recent years, specialist segmentation approaches are limited to specific tasks and data distribution. Retraining a new model for adaptation to new scenarios or settings takes expensive computation and time cost, which raises the demand for versatile and universal segmentation model that can cater to various granularity. Although some attempts have been made for unifying different segmentation tasks or generalization to various scenarios, limitations in the definition of paradigms and input-output spaces make it difficult for them to achieve accurate understanding of content at arbitrary granularity. To this end, we present UniLSeg, a universal segmentation model that can perform segmentation at any semantic level with the guidance of language instructions. For training UniLSeg, we reorganize a group of tasks from original diverse distributions into a unified data format, where images with texts describing segmentation targets as input and corresponding masks are output. Combined with a automatic annotation engine for utilizing numerous unlabeled data, UniLSeg achieves excellent performance on various tasks and settings, surpassing both specialist and unified segmentation models.

This paper proposes a cross-modal distillation framework, PartDistill, which transfers 2D knowledge from vision-language models (VLMs) to facilitate 3D shape part segmentation. PartDistill addresses three major challenges in this task: the lack of 3D segmentation in invisible or undetected regions in the 2D projections, inconsistent 2D predictions by VLMs, and the lack of knowledge accumulation across different 3D shapes. PartDistill consists of a teacher network that uses a VLM to make 2D predictions and a student network that learns from the 2D predictions while extracting geometrical features from multiple 3D shapes to carry out 3D part segmentation. A bi-directional distillation, including forward and backward distillations, is carried out within the framework, where the former forward distills the 2D predictions to the student network, and the latter improves the quality of the 2D predictions, which subsequently enhances the final 3D segmentation. Moreover, PartDistill can exploit generative models that facilitate effortless 3D shape creation for generating knowledge sources to be distilled. Through extensive experiments, PartDistill boosts the existing methods with substantial margins on widely used ShapeNetPart and PartNetE datasets, by more than 15\% and 12\% higher mIoU scores, respectively. The code for this work is available at https://github.com/ardianumam/PartDistill.

The creation of personalized anatomical digital twins is important in the fields of medicine, computer graphics, sports science, and biomechanics. To observe a subject's anatomy, expensive medical devices (MRI or CT) are required and the creation of the digital model is often time-consuming and involves manual effort. Instead, we leverage the fact that the shape of the body surface is correlated with the internal anatomy; e.g. from surface observations alone, one can predict body composition and skeletal structure. In this work, we go further and learn to infer the 3D location of three important anatomic tissues: subcutaneous adipose tissue (fat), lean tissue (muscles and organs), and long bones. To learn to infer these tissues, we tackle several key challenges. We first create a dataset of human tissues by segmenting full-body MRI scans and registering the SMPL body mesh to the body surface. With this dataset, we train HIT (Human Implicit Tissues), an implicit function that, given a point inside a body, predicts its tissue class. HIT leverages the SMPL body model shape and pose parameters to canonicalize the medical data. Unlike SMPL, which is trained from upright 3D scans, MRI scans are acquired with subjects lying on a table, resulting in significant soft-tissue deformation. Consequently, HIT uses a learned volumetric deformation field that undoes these deformations. Since HIT is parameterized by SMPL, we can repose bodies or change the shape of subjects and the internal structures deform appropriately. We perform extensive experiments to validate HIT's ability to predict a plausible internal structure for novel subjects. The dataset and HIT model are available at https://hit.is.tue.mpg.de to foster future research in this direction.

This paper studies open-vocabulary segmentation (OVS) through calibrating in-vocabulary and domain-biased embedding space with generalized contextual prior of CLIP. As the core of open-vocabulary understanding, alignment of visual content with the semantics of unbounded text has become the bottleneck of this field. To address this challenge, recent works propose to utilize CLIP as an additional classifier and aggregate model predictions with CLIP classification results. Despite their remarkable progress, performance of OVS methods in relevant scenarios is still unsatisfactory compared with supervised counterparts. We attribute this to the in-vocabulary embedding and domain-biased CLIP prediction. To this end, we present a Semantic-assisted CAlibration Network (SCAN). In SCAN, we incorporate generalized semantic prior of CLIP into proposal embedding to avoid collapsing on known categories. Besides, a contextual shift strategy is applied to mitigate the lack of global context and unnatural background noise.With above designs, SCAN achieves state-of-the-art performance on all popular open-vocabulary segmentation benchmarks. Furthermore, we also focus on the problem of existing evaluation system that ignores semantic duplication across categories, and propose a new metric called Semantic-Guided IoU (SG-IoU). Code is available here.

Interactive Segmentation (IS) segments specific objects or parts in the image according to user input. Current IS pipelines fall into two categories: single-granularity output and multi-granularity output. The latter aims to alleviate the spatial ambiguity present in the former. However, the multi-granularity output pipeline suffers from limited interaction flexibility and produces redundant results. In this work, we introduce Granularity-Controllable Interactive Segmentation (GraCo), a novel approach that allows precise control of prediction granularity by introducing additional parameters to input. This enhances the customization of the interactive system and eliminates redundancy while resolving ambiguity. Nevertheless, the exorbitant cost of annotating multi-granularity masks and the lack of available datasets with granularity annotations make it difficult for models to acquire the necessary guidance to control output granularity. To address this problem, we design an any-granularity mask generator that exploits the semantic property of the pre-trained IS model to automatically generate abundant mask-granularity pairs without requiring additional manual annotation. Based on these pairs, we propose a granularity-controllable learning strategy that efficiently imparts the granularity controllability to the IS model. Extensive experiments on intricate scenarios at object and part levels demonstrate that our GraCo has significant advantages over previous methods. This highlights the potential of GraCo to be a flexible annotation tool, capable of adapting to diverse segmentation scenarios. Code is available in the supplementary material.

The Segment Anything Model (SAM) has garnered significant attention for its versatile segmentation abilities and intuitive prompt-based interface. However, its application in medical imaging presents challenges, requiring either substantial training cost and extensive medical datasets for full model fine-tuning or high-quality prompts for optimal performance.This paper introduces H-SAM: a prompt-free adaptation of SAM designed for efficient fine-tuning on medical images via a two-stage hierarchical decoding procedure. In the first stage, H-SAM employs SAM's original decoder to create a prior (probabilistic) mask, which will be used to guide more intricate decoding in the second stage. Specifically, we propose two key designs: 1) A class-balanced, mask-guided self-attention mechanism that addresses the unbalanced label distribution and thus enhancing the image embedding; 2) A learnable mask cross-attention mechanism that spatially modulates the interplay among different image regions based on the prior mask. Moreover, the inclusion of a hierarchical pixel decoder in H-SAM enhances its proficiency in capturing fine-grained and localized details. This approach enables SAM to effectively integrate learned medical prior, facilitating enhanced adaptation for medical image segmentation with limited samples.Our H-SAM enjoys 4.78\% improvement in average Dice compared to existing prompt-free SAM variants for multi-organ segmentation using only 10\% 2D slices.Without using any unlabeled data at all, H-SAM is able to even outperform state-of-the-art semi-supervised models which use extensive unlabeled training data on various medical datasets.

Semantic segmentation has innately relied on extensive pixel-level annotated data, leading to the emergence of unsupervised methodologies. Among them, leveraging self-supervised Vision Transformers for unsupervised semantic segmentation (USS) has been making steady progress with expressive deep features. Yet, for semantically segmenting images with complex objects, a predominant challenge remains: the lack of explicit object-level semantic encoding in patch-level features. This technical limitation often leads to inadequate segmentation of complex objects with diverse structures. To address this gap, we present a novel approach, EAGLE, which emphasizes object-centric representation learning for unsupervised semantic segmentation. Specifically, we introduce EiCue, a spectral technique providing semantic and structural cues through an eigenbasis derived from the semantic similarity matrix of deep image features and color affinity from an image. Further, by incorporating our object-centric contrastive loss with EiCue, we guide our model to learn object-level representations with intra- and inter-image object-feature consistency, thereby enhancing semantic accuracy. Extensive experiments on COCO-Stuff, Cityscapes, and Potsdam-3 datasets demonstrate the state-of-the-art USS results of EAGLE with accurate and consistent semantic segmentation across complex scenes.

Recently, One-stage Weakly Supervised Semantic Segmentation (WSSS) with image-level labels has gained increasing interest due to simplification over its cumbersome multi-stage counterpart. Limited by the inherent ambiguity of Class Activation Map (CAM), we observe that one-stage pipelines often encounter confirmation bias caused by incorrect CAM pseudo-labels, impairing their final segmentation performance. Although recent works discard many unreliable pseudo-labels to implicitly alleviate this issue, they fail to exploit sufficient supervision for their models. To this end, we propose a dual student framework with trustworthy progressive learning (DuPL). Specifically, we propose a dual student network with a discrepancy loss to yield diverse CAMs for each sub-net. The two sub-nets generate supervision for each other, mitigating the confirmation bias caused by learning their own incorrect pseudo-labels. In this process, we progressively introduce more trustworthy pseudo-labels to be involved in the supervision through dynamic threshold adjustment with an adaptive noise filtering strategy. Moreover, we believe that every pixel, even discarded from supervision due to its unreliability, is important for WSSS. Thus, we develop consistency regularization on these discarded regions, providing supervision of every pixel. Experiment results demonstrate the superiority of the proposed DuPL over the recent state-of-the-art alternatives on PASCAL VOC 2012 and MS COCO datasets.

The estimation of implicit cross-frame correspondences and the high computational cost have long been major challenges in video semantic segmentation (VSS) for driving scenes. Prior works utilize keyframes, feature propagation, or cross-frame attention to address these issues.By contrast, we are the first to harness vanishing point (VP) priors for more effective segmentation. Intuitively, objects near VPs (i.e., away from the vehicle) are less discernible. Moreover, they tend to move radially away from the VP over time in the usual case of a forward-facing camera, a straight road, and linear forward motion of the vehicle. Our novel, efficient network for VSS, named VPSeg, incorporates two modules that utilize exactly this pair of static and dynamic VP priors: sparse-to-dense feature mining (DenseVP) and VP-guided motion fusion (MotionVP). MotionVP employs VP-guided motion estimation to establish explicit correspondences across frames and help attend to the most relevant features from neighboring frames, while DenseVP enhances weak dynamic features in distant regions around VPs. These modules operate within a context-detail framework, which separates contextual features from high-resolution local features at different input resolutions to reduce computational costs. Contextual and local features are integrated through contextualized motion attention (CMA) for the final prediction. Extensive experiments on two popular driving segmentation benchmarks, Cityscapes and ACDC, demonstrate that VPSeg outperforms previous SOTA methods, with only modest computational overhead.

Producing quality segmentation masks for images is a fundamental problem in computer vision. Recent research has explored large-scale supervised training to enable zero-shot segmentation on virtually any image style and unsupervised training to enable segmentation without dense annotations. However, constructing a model capable of segmenting anything in a zero-shot manner without any annotations is still challenging. In this paper, we propose to utilize the self-attention layers in stable diffusion models to achieve this goal because the pre-trained stable diffusion model has learned inherent concepts of objects within its attention layers. Specifically, we introduce a simple yet effective iterative merging process based on measuring KL divergence among attention maps to merge them into valid segmentation masks. The proposed method does not require any training or language dependency to extract quality segmentation for any images. On COCO-Stuff-27, our method surpasses the prior unsupervised zero-shot SOTA method by an absolute $26\%$ in pixel accuracy and $17\%$ in mean IoU.

State-of-the-art models on contemporary 3D segmentation benchmarks like ScanNet consume and label dataset-provided 3D point clouds, obtained through post processing of sensed multiview RGB-D images. They are typically trained in-domain, forego large-scale 2D pre-training and outperform alternatives that featurize the posed RGB-D multiview images instead. The gap in performance between methods that consume posed images versus post-processed 3D point clouds has fueled the belief that 2D and 3D perception require distinct model architectures. In this paper, we challenge this view and propose ODIN (Omni-Dimensional INstance segmentation), a model that can segment and label both 2D RGB images and 3D point clouds, using a transformer architecture that alternates between 2D within-view and 3D cross-view information fusion. Our model differentiates 2D and 3D feature operations through the positional encodings of the tokens involved, which capture pixel coordinates for 2D patch tokens and 3D coordinates for 3D feature tokens. ODIN achieves state-of-the-art performance on ScanNet200, Matterport3D and AI2THOR 3D instance segmentation benchmarks, and competitive performance on ScanNet, S3DIS and COCO. It outperforms all previous works by a wide margin when the sensed 3D point cloud is used in place of the point cloud sampled from 3D mesh. When used as the 3D perception engine in an instructable embodied agent architecture, it sets a new state-of-the-art on the TEACh action-from-dialogue benchmark. Our code and checkpoints can be found at the project website: https://odin-seg.github.io.

Due to high expense of human labor, one major challenge for semantic segmentation in real-world scenarios is the lack of sufficient pixel-level labels, which is more serious when processing video data. To exploit unlabeled data for model training, semi-supervised learning methods attempt to construct pseudo labels or various auxiliary constraints as supervision signals. However, most of them just process video data as a set of independent images in a per-frame manner. The rich temporal relationships are ignored, which can serve as valuable clues for representation learning. Besides, this per-frame recognition paradigm is quite different from that of humans. Actually, benefited from the internal temporal relevance of video data, human would wisely use the distinguished semantic concepts in historical frames to aid the recognition of the current frame. Motivated by this observation, we propose a novel temporally-dependent classifier (TDC) to mimic the human-like recognition procedure. Comparing to the conventional classifier, TDC can guide the model to learn a group of temporally-consistent semantic concepts across frames, which essentially provides an implicit and effective constraint. We conduct extensive experiments on Cityscapes and CamVid, and the results demonstrate the superiority of our proposed method to previous state-of-the-art methods.

Single-point annotation in visual tasks, with the goal of minimizing labelling costs, is becoming increasingly prominent in research. Recently, visual foundation models, such as Segment Anything (SAM), have gained widespread usage due to their robust zero-shot capabilities and exceptional annotation performance. However, SAM's class-agnostic output and high confidence in local segmentation introduce 'semantic ambiguity', posing a challenge for precise category-specific segmentation. In this paper, we introduce a cost-effective category-specific segmenter using SAM. To tackle this challenge, we have devised a Semantic-Aware Instance Segmentation Network (SAPNet) that integrates Multiple Instance Learning (MIL) with matching capability and SAM with point prompts. SAPNet strategically selects the most representative mask proposals generated by SAM to supervise segmentation, with a specific focus on object category information. Moreover, we introduce the Point Distance Guidance and Box Mining Strategy to mitigate inherent challenges: 'group' and 'local' issues in weakly supervised segmentation. These strategies serve to further enhance the overall segmentation performance. The experimental results on Pascal VOC and COCO demonstrate the promising performance of our proposed SAPNet, emphasizing its semantic matching capabilities and its potential to advance point-prompted instance segmentation. The code are available at https://github.com/CVPR666/SAPNet.

Weakly Supervised Semantic Segmentation (WSSS) relies on Class Activation Maps (CAMs) to extract spatial information from image-level labels. With the success of Vision Transformer (ViT), the migration of ViT is actively conducted in WSSS. This work proposes a novel WSSS framework with Class Token Infusion (CTI). By infusing the class tokens from images, we guide class tokens to possess class-specific distinct characteristics and global-local consistency. For this, we devise two kinds of token infusion: 1) Intra-image Class Token Infusion (I-CTI) and 2) Cross-Image Class Token Infusion (C-CTI). In I-CTI, we infuse the class tokens from the same but differently augmented images and thus make CAMs consistent among various deformations (i.e., view, color). In C-CTI, by infusing the class tokens from the other images and imposing the resulting CAMs to be similar, it learns class-specific distinct characteristics. Besides the CTI, we bring the background (BG) concept into ViT with the BG token to reduce the false positive activation of CAMs. We demonstrate the effectiveness of our method on PASCAL VOC 2012 and MS COCO 2014 datasets, achieving state-of-the-art results in weakly supervised semantic segmentation.

Weakly supervised semantic segmentation (WSSS) with image-level labels aims to achieve segmentation task without dense annotations. However, attributed to the frequent coupling of co-occurring objects and the limited supervision from image-level labels, the challenging co-occurrence problem is widely present and leads to false activation of objects. In this work, we devise a 'Separate and Conquer' training paradigm SeCo to tackle the co-occurrence challenge from dimensions of image space and feature space. In the image space, we propose to 'separate' the co-occurring objects with image decomposition by subdividing images into patches. Importantly, we assign each patch a category tag from Class Activation Maps (CAMs), which helps to identify objects at patch level and guide the subsequent representation. In the feature space, we propose to 'conquer' the false activation by enhancing semantic representation with multi-granularity knowledge contrast. To this end, a dual-teacher-single-student architecture is designed to extract knowledge at image level and patch level. Along with the knowledge and patch tags, class-specific contrast is conducted to facilitate the discrepancy among co-occurring objects. We streamline the multi-staged WSSS pipeline end-to-end and tackle co-occurrence without external supervision. Extensive experiments are conducted, validating the efficiency of our method tackling co-occurrence and the superiority over previous single-staged and even multi-staged competitors on PASCAL VOC and MS COCO. Codes will be publicly available soon.

Deep learning models for semantic segmentation often experience performance degradation when deployed to unseen target domains unidentified during the training phase. This is mainly due to variations in image texture (\ie style) from different data sources. To tackle this challenge, existing domain generalized semantic segmentation (DGSS) methods attempt to remove style variations from the feature. However, these approaches struggle with the entanglement of style and content, which may lead to the unintentional removal of crucial content information, causing performance degradation. This study addresses this limitation by proposing BlindNet, a novel DGSS approach that blinds the style without external modules or datasets. The main idea behind our proposed approach is to alleviate the effect of style in the encoder whilst facilitating robust segmentation in the decoder. To achieve this, BlindNet comprises two key components: covariance alignment and semantic consistency contrastive learning. Specifically, the covariance alignment trains the encoder to uniformly recognize various styles and preserve the content information of the feature, rather than removing the style-sensitive factor. Meanwhile, semantic consistency contrastive learning enables the decoder to construct discriminative class embedding space and disentangles features that are vulnerable to misclassification. Through extensive experiments, our approach outperforms existing DGSS methods, exhibiting robustness and superior performance for semantic segmentation on unseen target domains.

Semi-supervised semantic segmentation (SSSS) has been proposed to alleviate the burden of time-consuming pixel-level manual labeling, which leverages limited labeled data along with larger amounts of unlabeled data. Current state-of-the-art methods train the labeled data with ground truths and unlabeled data with pseudo labels. However, the two training flows are separate, which allows labeled data to dominate the training process, resulting in low-quality pseudo labels and, consequently, sub-optimal results. To alleviate this issue, we present AllSpark, which reborns the labeled features from unlabeled ones with the channel-wise cross-attention mechanism. We further introduce a Semantic Memory along with a Channel Semantic Grouping strategy to ensure that unlabeled features adequately represent labeled features. The AllSpark shed new light on the architecture level designs of SSSS rather than framework level, which avoids increasingly complicated training pipeline designs. It can also be regarded as a flexible bottleneck module that can be seamlessly integrated into a general transformer-based segmentation model. The proposed AllSpark outperforms existing methods across all evaluation protocols on Pascal, Cityscapes and COCO benchmarks without bells-and-whistles. Code and model weights are available at: https://github.com/xmed-lab/AllSpark.

Traditionally, training neural networks to perform semantic segmentation required expensive human-made annotations. But more recently, advances in the field of unsupervised learning have made significant progress on this issue and towards closing the gap to supervised algorithms. To achieve this, semantic knowledge is distilled by learning to correlate randomly sampled features from images across an entire dataset. In this work, we build upon these advances by incorporating information about the structure of the scene into the training process through the use of depth information. We achieve this by (1) learning depth-feature correlation by spatially correlating the feature maps with the depth maps to induce knowledge about the structure of the scene and (2) exploiting farthest-point sampling to more effectively select relevant features by utilizing 3D sampling techniques on depth information of the scene. Finally, we demonstrate the effectiveness of our technical contributions through extensive experimentation and present significant improvements in performance across multiple benchmark datasets.

Although polygon meshes have been a standard representation in geometry processing, their irregular and combinatorial nature hinders their suitability for learning-based applications. In this paper we introduce a novel learnable mesh representation through a set of local 3D sample Points and their associated Normals and Quadric error metrics (QEM) w.r.t. the underlying shape, which we denote PoNQ. A global mesh is directly derived from PoNQ by efficiently leveraging the knowledge of the local quadric errors. Besides marking the first use of QEM within a neural shape representation, our contribution guarantees both topological and geometrical properties by ensuring that a PoNQ mesh does not self-intersect and is always the boundary of a volume. Notably, our representation does not rely on a regular grid, is supervised directly by the target surface alone, and also handles open surfaces with boundaries and/or sharp features. We demonstrate the efficacy of PoNQ through a learning-based mesh prediction from SDF grids and show that our method surpasses recent state-of-the-art techniques in terms of both surface and edge-based metrics.

Although 3D shape matching and interpolation are highly interrelated, they are often studied separately and applied sequentially to relate different 3D shapes, thus resulting in sub-optimal performance. In this work we present a unified framework to predict both point-wise correspondences and shape interpolation between 3D shapes. To this end, we combine the deep functional map framework with classical surface deformation models to map shapes in both spectral and spatial domains. On the one hand, by incorporating spatial maps, our method obtains more accurate and smooth point-wise correspondences compared to previous functional map methods for shape matching. On the other hand, by introducing spectral maps, our method gets rid of commonly used but computationally expensive geodesic distance constraints that are only valid for near-isometric shape deformations. Furthermore, we propose a novel test-time adaptation scheme to capture both pose-dominant and shape-dominant deformations. Using different challenging datasets, we demonstrate that our method outperforms previous state-of-the-art methods for both shape matching and interpolation, even compared to supervised approaches.

Co-salient object detection (CoSOD) aims to identify the common and salient (usually in the foreground) regions across a given group of images. Although achieving significant progress, state-of-the-art CoSODs could be easily affected by some adversarial perturbations, leading to substantial accuracy reduction. The adversarial perturbations can mislead CoSODs but do not change the high-level semantic information (e.g., concept) of the co-salient objects. In this paper, we propose a novel robustness enhancement framework by first learning the concept of the co-salient objects based on the input group images and then leveraging this concept to purify adversarial perturbations, which are subsequently fed to CoSODs for robustness enhancement. Specifically, we propose CosalPure containing two modules, i.e., group-image concept learning and concept-guided diffusion purification. For the first module, we adopt a pre-trained text-to-image diffusion model to learn the concept of co-salient objects within group images where the learned concept is robust to adversarial examples. For the second module, we map the adversarial image to the latent space and then perform diffusion generation by embedding the learned concept into the noise prediction function as an extra condition. Our method can effectively alleviate the influence of the SOTA adversarial attack containing different adversarial patterns, including exposure and noise. The extensive results demonstrate that our method could enhance the robustness of CoSODs significantly.

Sketch semantic segmentation is a well-explored and pivotal problem in computer vision involving the assignment of pre-defined part labels to individual strokes. This paper presents ContextSeg - a simple yet highly effective approach to tackling this problem with two stages. In the first stage, to better encode the shape and positional information of strokes, we propose to predict an extra dense distance field in an autoencoder network to reinforce structural information learning. In the second stage, we treat an entire stroke as a single entity and label a group of strokes within the same semantic part using an auto-regressive Transformer with the default attention mechanism. By group-based labeling, our method can fully leverage the context information when making decisions for the remaining groups of strokes. Our method achieves the best segmentation accuracy compared with state-of-the-art approaches on two representative datasets and has been extensively evaluated demonstrating its superior performance. Additionally, we offer insights into solving part imbalance in training data and the preliminary experiment on cross-category training, which can inspire future research in this field.

Open-vocabulary semantic segmentation aims at segmenting arbitrary categories expressed in textual form. Previous works have trained over large amounts of image-caption pairs to enforce pixel-level multimodal alignments. However, captions provide global information about the semantics of a given image but lack direct localization of individual concepts. Further, training on large-scale datasets inevitably brings significant computational costs. In this paper we propose FreeDA, a training-free diffusion-augmented method for open-vocabulary semantic segmentation, which leverages the ability of diffusion models to visually localize generated concepts and local-global similarities to match class-agnostic regions with semantic classes. Our approach involves an offline stage in which textual-visual reference embeddings are collected, starting from a large set of captions and leveraging visual and semantic contexts. At test time, these are queried to support the visual matching process, which is carried out by jointly considering class-agnostic regions and global semantic similarities. Extensive analyses demonstrate that FreeDA achieves state-of-the-art performance on five datasets, surpassing previous methods by more than 7.0 average points in terms of mIoU and without requiring any training. Our source code and models will be made publicly available.

In the evolving landscape of computer vision, foundation models have emerged as pivotal tools, exhibiting exceptional adaptability to a myriad of tasks. Among these, the Segment Anything Model (SAM) by Meta AI has distinguished itself in image segmentation. However, SAM, like its counterparts, encounters limitations in specific niche applications, prompting a quest for enhancement strategies that do not compromise its inherent capabilities. This paper introduces ASAM, a novel methodology that amplifies SAM's performance through adversarial tuning. We harness the potential of natural adversarial examples, inspired by their successful implementation in natural language processing. By utilizing a stable diffusion model, we augment a subset (1\%) of the SA-1B dataset, generating adversarial instances that are more representative of natural variations rather than conventional imperceptible perturbations. Our approach maintains the photorealism of adversarial examples and ensures alignment with original mask annotations, thereby preserving the integrity of the segmentation task. The fine-tuned ASAM demonstrates significant improvements across a diverse range of segmentation tasks without necessitating additional data or architectural modifications. The results of our extensive evaluations confirm that ASAM establishes new benchmarks in segmentation tasks, thereby contributing to the advancement of foundational models in computer vision. Our project page is in https://asam2024.github.io/.

We introduce in-context matting, a novel task setting of image matting. Given a reference image of a certain foreground and guided priors such as points, scribbles, and masks, in-context matting enables automatic alpha estimation on a batch of target images of the same foreground category, without additional auxiliary input. This setting marries good performance in auxiliary input-based matting and ease of use in automatic matting, which finds a good trade-off between customization and automation. To overcome the key challenge of accurate foreground matching, we introduce IconMatting, an in-context matting model built upon a pre-trained text-to-image diffusion model. Conditioned on inter- and intra-similarity matching, IconMatting can make full use of reference context to generate accurate target alpha mattes. To benchmark the task, we also introduce a novel testing dataset ICM-$57$, covering $57$ groups of real-world images. Quantitative and qualitative results on the ICM-$57$ testing set show that IconMatting rivals the accuracy of trimap-based matting while retaining the automation level akin to automatic matting. Code is available at https://github.com/tiny-smart/in-context-matting.

Manual annotation of every point in a point cloud is a costly and labor-intensive process. While weakly supervised point cloud semantic segmentation (WSPCSS) with sparse annotation shows promise, the limited information from initial sparse labels can place an upper bound on performance. As a new research direction for WSPCSS, we propose a novel Region Exploration via Artificial Labeling (REAL) framework. It leverages a foundational image model as an artificial oracle within the active learning context, eliminating the need for manual annotation by a human oracle. To integrate the 2D model into the 3D domain, we first introduce a Projection-based Point-to-Segment (PP2S) module, designed to enable prompt segmentation of 3D data without additional training. The REAL framework samples query points based on model predictions and requests annotations from PP2S, dynamically refining labels and improving model training. Furthermore, to overcome several challenges of employing an artificial model as an oracle, we formulate effective query sampling and label updating strategies. Our comprehensive experiments and comparisons demonstrate that the REAL framework significantly outperforms existing methods across various benchmarks. The code is available at https://github.com/jihun1998/AO.

Despite great improvements in semantic segmentation, challenges persist because of the lack of local/global contexts and the relationship between them. In this paper, we propose Contextrast, a contrastive learning-based semantic segmentation method that allows to capture local/global contexts and comprehend their relationships. Our proposed method comprises two parts: a) contextual contrastive learning (CCL) and b) boundary-aware negative (BANE) sampling. Contextual contrastive learning obtains local/global context from multi-scale feature aggregation and inter/intra-relationship of features for better discrimination capabilities. Meanwhile, BANE sampling selects embedding features along the boundaries of incorrectly predicted regions to employ them as harder negative samples on our contrastive learning, resolving segmentation issues along the boundary region by exploiting fine-grained details. We demonstrate that our Contextrast substantially enhances the performance of semantic segmentation networks, outperforming state-of-the-art contrastive learning approaches on diverse public datasets, e.g. Cityscapes, CamVid, PASCAL-C, COCO-Stuff, and ADE20K, without an increase in computational cost during inference.

Parameter-efficient fine-tuning (PEFT) is an effective methodology to unleash the potential of large foundation models in novel scenarios with limited training data. In the computer vision community, PEFT has shown effectiveness in image classification, but little research has studied its ability for image segmentation. Fine-tuning segmentation models usually require a heavier adjustment of parameters to align the proper projection directions in the parameter space for new scenarios. This raises a challenge to existing PEFT algorithms, as they often inject a limited number of individual parameters into each block, which prevents substantial adjustment of the projection direction of the parameter space due to the limitation of Hidden Markov Chain along blocks. In this paper, we equip PEFT with a cross-block orchestration mechanism to enable the adaptation of the Segment Anything Model (SAM) to various downstream scenarios. We introduce a novel inter-block communication module, which integrates a learnable relation matrix to facilitate communication among different coefficient sets of each PEFT block’s parameter space. Moreover, we propose an intra-block enhancement module, which introduces a linear projection head whose weights are generated from a hyper-complex layer, further enhancing the impact of the adjustment of projection directions on the entire parameter space. Extensive experiments on diverse benchmarks demonstrate that our proposed approach consistently improves the segmentation performance significantly on novel scenarios with only around 1K** additional parameters.

CAD programs are a popular way to compactly encode shapes as a sequence of operations that are easy to parametrically modify. However, without sufficient semantic comments and structure, such programs can be challenging to understand, let alone modify. We introduce the problem of semantic commenting CAD programs, wherein the goal is to segment the input program into code blocks corresponding to semantically meaningful shape parts and assign a semantic label to each block. We solve the problem by combining program parsing with visual-semantic analysis afforded by recent advances in foundational language and vision models. Specifically, by executing the input programs, we create shapes, which we use to generate conditional photorealistic images to make use of semantic annotators for such images. We then distill the information across the images and link back to the original programs to semantically comment on them. Additionally, we collected and annotated a benchmark dataset, CADTalk, consisting of 5,288 machine-made programs and 45 human-made programs with ground truth semantic comments. We extensively evaluated our approach, compared it to a GPT-based baseline, and an open-set shape segmentation baseline, and reported an 83.24% accuracy on the new CADTalk dataset. Code and data: https://enigma-li.github.io/CADTalk/.

Computer-Aided Design (CAD) model reconstruction from point clouds is an important problem at the intersection of computer vision, graphics, and machine learning; it saves the designer significant time when iterating on in-the-wild objects. Recent advancements in this direction achieve relatively reliable semantic segmentation but still struggle to produce an adequate topology of the CAD model. In this work, we analyze the current state of the art for that ill-posed task and identify shortcomings of existing methods. We propose a hybrid analytic-neural reconstruction scheme that bridges the gap between segmented point clouds and structured CAD models and can be readily combined with different segmentation backbones. Moreover, to power the surface fitting stage, we propose a novel implicit neural representation of freeform surfaces, driving up the performance of our overall CAD reconstruction scheme. We extensively evaluate our method on the popular ABC benchmark of CAD models and set a new state-of-the-art for that dataset. Code and models will be shared publicly.

The goal of interactive image segmentation is to delineate specific regions within an image via visual or language prompts. Low-latency and high-quality interactive segmentation with diverse prompts remain challenging for existing specialist and generalist models. Specialist models, with their limited prompts and task-specific designs, experience high latency because the image must be recomputed every time the prompt is updated, due to the joint encoding of image and visual prompts. Generalist models, exemplified by the Segment Anything Model (SAM), have recently excelled in prompt diversity and efficiency, lifting image segmentation to the foundation model era. However, for high-quality segmentations, SAM still lags behind state-of-the-art specialist models despite SAM being trained with x100 more segmentation masks. In this work, we delve deep into the architectural differences between the two types of models. We observe that dense representation and fusion of visual prompts are the key design choices contributing to the high segmentation quality of specialist models. In light of this, we reintroduce this dense design into the generalist models, to facilitate the development of generalist models with high segmentation quality. To densely represent diverse visual prompts, we propose to use a dense map to capture five types: clicks, boxes, polygons, scribbles, and masks. Thus, we propose SegNext, a next-generation interactive segmentation approach offering low latency, high quality, and diverse prompt support. Our method outperforms current state-of-the-art methods on HQSeg-44K and DAVIS quantitatively and qualitatively.

We present GLEE in this work, an object-level foundation model for locating and identifying objects in images and videos. Through a unified framework, GLEEaccomplishes detection, segmentation, tracking, grounding, and identification of arbitrary objects in the open world scenario for various object perception tasks. Adopting a cohesive learning strategy, GLEE acquires knowledge from diverse data sources with varying supervision levels to formulate general object representations, excelling in zero-shot transfer to new data and tasks. Specifically, we employ an image encoder, text encoder, and visual prompter to handle multi-modal inputs, enabling to simultaneously solve various object-centric downstream tasks while maintaining state-of-the-art performance. Demonstrated through extensive training on over five million images from diverse benchmarks, GLEE exhibits remarkable versatility and improved generalization performance, efficiently tackling downstream tasks without the need for task-specific adaptation. By integrating large volumes of automatically labeled data, we further enhance its zero-shot generalization capabilities. Additionally, GLEE is capable of being integrated into Large Language Models, serving as a foundational model to provide universal object-level information for multi-modal tasks. We hope that the versatility and universality of our method will mark a significant step in the development of efficient visual foundation models for AGI systems. The models and code are released at https://github.com/FoundationVision/GLEE.

Weakly supervised semantic segmentation has witnessed great achievements with image-level labels. Several recent approaches use the CLIP model to generate pseudo labels for training an individual segmentation model, while there is no attempt to apply the CLIP model as the backbone to directly segment objects with image-level labels. In this paper, we propose WeCLIP, a CLIP-based single-stage pipeline, for weakly supervised semantic segmentation. Specifically, the frozen CLIP model is applied as the backbone for semantic feature extraction, and a new decoder is designed to interpret extracted semantic features for final prediction. Meanwhile, we utilize the above frozen backbone to generate pseudo labels for training the decoder. Such labels cannot be optimized during training. We then propose a refinement module (RFM) to rectify them dynamically. Our architecture enforces the proposed decoder and RFM to benefit from each other to boost the final performance. Extensive experiments show that our approach significantly outperforms other approaches with less training cost. Additionally, our WeCLIP also obtains promising results for fully supervised settings. The code is available at https://github.com/zbf1991/WeCLIP.

Unsupervised video object segmentation aims to segment the most prominent object in a video sequence. However, the existence of complex backgrounds and multiple foreground objects make this task challenging. To address this issue, we propose a guided slot attention network to reinforce spatial structural information and obtain better foreground-background separation. The foreground and background slots, which are initialized with query guidance, are iteratively refined based on interactions with template information. Furthermore, to improve slot--template interaction and effectively fuse global and local features in the target and reference frames, K-nearest neighbors filtering and a feature aggregation transformer are introduced. The proposed model achieves state-of-the-art performance on two popular datasets. Additionally, we demonstrate the robustness of the proposed model in challenging scenes through various comparative experiments.

The recent advent of pre-trained vision transformers has unveiled a promising property: their inherent capability to group semantically related visual concepts. In this paper, we explore to harnesses this emergent feature to tackle few-shot semantic segmentation, a task focused on classifying pixels in a test image with a few example data. A critical hurdle in this endeavor is preventing overfitting to the limited classes seen during training the few-shot segmentation model. As our main discovery, we find that the concept of "relationship descriptors", initially conceived for enhancing the CLIP model for zero-shot semantic segmentation, offers a potential solution. We adapt and refine this concept to craft a relationship descriptor construction tailored for few-shot semantic segmentation, extending its application across multiple layers to enhance performance. Building upon this adaptation, we proposed a few-shot semantic segmentation framework that is not only easy to implement and train but also effectively scales with the number of support examples and categories. Through rigorous experimentation across various datasets, including PASCAL-$5^{i}$ and COCO-$20^{i}$, we demonstrate a clear advantage of our method in diverse few-shot semantic segmentation scenarios, and a range of pre-trained vision transformer models. The findings clearly show that our method significantly outperforms current state-of-the-art techniques, highlighting the effectiveness of harnessing the emerging capabilities of vision transformers for few-shot semantic segmentation.

Vision-language foundation models have shown remarkable performance in various zero-shot settings such as image retrieval, classification, or captioning. But so far, those models seem to fall behind when it comes to zero-shot localization of referential expressions and objects in images. As a result, they need to be fine-tuned for this task.In this paper we show that pretrained vision-language (VL) models allow for zero-shot open-vocabulary object localization without any fine-tuning. To leverage those capabilities, we propose a Grounding Everything Module (GEM) that generalizes the idea of value-value attention introduced by CLIPSurgery to a self-self attention path. We show that the concept of self-self attention corresponds to clustering, thus enforcing groups of tokens arising from the same object to be similar while preserving the alignment with the language space. To further guide the group formation, we propose a set of regularizations that allows the model to finally generalize across datasets and backbones. We evaluate the proposed GEM framework on various benchmark tasks and datasets for semantic segmentation. It shows that GEM not only outperforms other training-free open-vocabulary localization methods, but also achieves state-of-the-art results on the recently proposed OpenImagesV7 large-scale segmentation benchmark.

To reduce the reliance on large-scale datasets, recent works in 3D segmentation resort to few-shot learning. Current 3D few-shot segmentation methods first pre-train models on 'seen' classes, and then evaluate their generalization performance on 'unseen' classes. However, the prior pre-training stage not only introduces excessive time overhead but also incurs a significant domain gap on 'unseen' classes. To tackle these issues, we propose a Non-parametric Network for few-shot 3D Segmentation, Seg-NN, and its Parameterized variant, Seg-PN. Without training, Seg-NN extracts dense representations by hand-crafted filters and achieves comparable performance to existing parameterized models. Due to the elimination of pre-training, Seg-NN can alleviate the domain gap issue and save a substantial amount of time. Based on Seg-NN, Seg-PN only requires training a lightweight QUEry-Support Transferring (QUEST) module, which enhances the interaction between the support set and query set. Experiments suggest that Seg-PN outperforms previous state-of-the-art method by +4.19% and +7.71% mIoU on S3DIS and ScanNet datasets respectively, while reducing training time by -90%, indicating its effectiveness and efficiency.

Most continual segmentation methods tackle the problem as a per-pixel classification task. However, such a paradigm is very challenging, and we find query-based segmenters with built-in objectness have inherent advantages compared with per-pixel ones, as objectness has strong transfer ability and forgetting resistance. Based on these findings, we propose CoMasTRe by disentangling continual segmentation into two stages: forgetting-resistant continual objectness learning and well-researched continual classification. CoMasTRe uses a two-stage segmenter learning class-agnostic mask proposals at the first stage and leaving recognition to the second stage. During continual learning, a simple but effective distillation is adopted to strengthen objectness. To further mitigate the forgetting of old classes, we design a multi-label class distillation strategy suited for segmentation. We assess the effectiveness of CoMasTRe on PASCAL VOC and ADE20K. Extensive experiments show that our method outperforms per-pixel and query-based methods on both datasets. Code will be available at https://github.com/jordangong/CoMasTRe.

Generalized Referring Expression Segmentation (GRES) extends the scope of classic RES to refer to multiple ob-jects in one expression or identify the empty targets absent in the image. GRES poses challenges in modeling the com-plex spatial relationships of the instances in the image and identifying non-existing referents. Multimodal Large Lan-guage Models (MLLMs) have recently shown tremendous progress in these complicated vision-language tasks. Con-necting Large Language Models (LLMs) and vision models, MLLMs are proficient in understanding contexts with visual inputs. Among them, LISA, as a representative, adopts a special [SEG] token to prompt a segmentation mask de-coder, e.g., SAM, to enable MLLMs in the RES task. How-ever, existing solutions to GRES remain unsatisfactory since current segmentation MLLMs cannot correctly handle the cases where users might reference multiple subjects in a singular prompt or provide descriptions incongruent with any image target. In this paper, we propose Generalized Segmentation Vision Assistant (GSVA) to address this gap. Specifically, GSVA reuses the [SEG] token to prompt the segmentation model towards supporting multiple mask ref-erences simultaneously and innovatively learns to generate a [REJ] token to reject the null targets explicitly. Experi-ments validate GSVA’s efficacy in resolving the GRES issue, marking a notable enhancement and setting a new record on the GRES benchmark gRefCOCO dataset. GSVA also proves effective across various classic referring segmenta-tion and comprehension tasks. Code will be available at https://github.com/LeapLabTHU/GSVA.

Human matting is a foundation task in image and video processing where human foreground pixels are extracted from the input. Prior works either improve the accuracy by additional guidance or improve the temporal consistency of a single instance across frames. We propose a new framework MaGGIe, Masked Guided Gradual Human Instance Matting, which predicts alpha mattes progressively for each human instances while maintaining the computational cost, precision, and consistency. Our method leverages modern architectures, including transformer attention and sparse convolution, to output all instance mattes simultaneously without exploding memory and latency. Although keeping constant inference costs in the multiple-instance scenario, our framework achieves robust and versatile performance on our proposed synthesized benchmarks. With the higher quality image and video matting benchmarks, the novel multi-instance synthesis approach from publicly available sources is introduced to increase the generalization of models in real-world scenarios. Our code and datasets are available at https://maggie-matt.github.io

The portrait matting task aims to extract an alpha matte with complete semantics and finely detailed contours. In comparison to CNN-based approaches, transformers with self-attention module have a better capacity to capture long-range dependencies and low-frequency semantic information of a portrait.However, recent research shows that the self-attention mechanism struggles with modeling high-frequency contour information and capturing fine contour details, which can lead to bias while predicting the portrait's contours. To deal with this issue, we propose EFormer to enhance the model's attention towards both the low-frequency semantic and high-frequency contour features. For the high-frequency contours, our research demonstrates that cross-attention module between different resolutions can guide our model to allocate attention appropriately to these contour regions.Supported by this, we can successfully extract the high-frequency detail information around the portrait's contours, which were previously ignored by self-attention.Based on the cross-attention module, we further build a semantic and contour detector (SCD) to accurately capture both the low-frequency semantic and high-frequency contour features.And we design a contour-edge extraction branch and semantic extraction branch to extract refined high-frequency contour features and complete low-frequency semantic information, respectively.Finally, we fuse the two kinds of features and leverage the segmentation head to generate a predicted portrait matte. Experiments on VideoMatte240K (JPEG SD Format) and Adobe Image Matting (AIM) datasets demonstrate that EFormer outperforms previous portrait matte methods.

In this paper, we delve into the nuanced challenge of tailoring the Segment Anything Models (SAMs) for integration with event data, with the overarching objective of attaining robust and universal object segmentation within the event-centric domain. One pivotal issue at the heart of this endeavor is the precise alignment and calibration of embeddings derived from event-centric data such that they harmoniously coincide with those originating from RGB imagery. Capitalizing on the vast repositories of datasets with paired events and RGB images, our proposition is to harness and extrapolate the profound knowledge encapsulated within the pre-trained SAM framework. As a cornerstone to achieving this, we introduce a multi-scale feature distillation methodology. This methodology rigorously optimizes the alignment of token embeddings originating from event data with their RGB image counterparts, thereby preserving and enhancing the robustness of the overall architecture. Considering the distinct significance that token embeddings from intermediate layers hold for higher-level embeddings, our strategy is centered on accurately calibrating the pivotal token embeddings. This targeted calibration is aimed at effectively managing the discrepancies in high-level embeddings originating from both the event and image domains. Extensive experiments on different datasets demonstrate the effectiveness of the proposed distillation method. Code in https://github.com/happychenpipi/EventSAM.

The creation of high-quality alpha mattes as ground-truth data for video matting is typically a laborious task. The trade-off between accuracy, manual corrections, and capture constraints often produces erroneous results or is cost prohibitive. We propose PolarMatte, a fully computational alpha matte extraction method for images and video without compromise between quality and practicality. A single polarization camera is used to capture dynamic scenes backlit by an off-the-shelf LCD monitor. PolarMatte exploits the polarization channel to compute the per-pixel opacity of the target scene, including the transparency of fine-details, translucent objects, and optical/motion blur. We leverage polarization clues to robustly detect indistinguishable pixels, and extract the alpha matte value at polarized foreground reflections with a polarimetric matting Laplacian. Quantitative and qualitative evaluation demonstrate our ability to computationally extract ground-truth-quality alpha mattes without human labour.

Semantic segmentation models, while effective for in-distribution categories, face challenges in real-world deployment due to encountering out-of-distribution (OoD) objects. Detecting these OoD objects is crucial for safety-critical applications. Existing methods rely on anomaly scores, but choosing a suitable threshold for generating masks presents difficulties and can lead to fragmentation and inaccuracy. This paper introduces a method to convert anomaly $\underline{S}$core $\underline{T}$o segmentation $\underline{M}$ask, called S2M, a simple and effective framework for OoD detection in semantic segmentation. Unlike assigning anomaly scores to pixels, S2M directly segments the entire OoD object. By transforming anomaly scores into prompts for a promptable segmentation model, S2M eliminates the need for threshold selection. Extensive experiments demonstrate that S2M outperforms the state-of-the-art by approximately 20\% in IoU and 40\% in mean F1 score, on average, across various benchmarks including Fishyscapes, Segment-Me-If-You-Can, and RoadAnomaly datasets.

Dichotomous Image Segmentation (DIS) has recently emerged towards highly accurate objects from high-resolution natural images.When designing an effective DIS model, the most challenge is how to balance the semantic dispersion of high-resolution targets in the small receptive field and the loss of high-precision details in the large receptive field. Existing methods rely on tedious multiple encoder-decoder streams and stages to gradually complete the global localization and local refinement. Inspired by the human visual system capturing regions of interest by observing from multiple views, we model DIS as a multi-view object perception problem and provide a parsimonious multi-view aggregation network (MVANet), which unifies the feature fusion of the distant view and close-up view into a single stream with one encoder-decoder structure. With the help of the proposed multi-view complementary localization and refinement modules, our approach established long-range, profound visual interactions across multiple views, allowing the features of the detailed close-up view to focus on refining highly accurate details. Experiments on the popular DIS-5K dataset show that our MVANet significantly outperforms state-of-the-art methods in both accuracy and speed.

We introduce pix2gestalt, a framework for zero-shot amodal segmentation, which learns to estimate the shape and appearance of whole objects that are only partially visible behind occlusions. By capitalizing on large-scale diffusion models and transferring their representations to this task, we learn a conditional diffusion model for reconstructing whole objects in challenging zero-shot cases, including examples that break natural and physical priors, such as art. As training data, we use a synthetically curated dataset containing occluded objects paired with their whole counterparts. Experiments show that our approach outperforms supervised baselines on established benchmarks. Our model can furthermore be used to significantly improve the performance of existing object recognition and 3D reconstruction methods in the presence of occlusions.

Few-shot segmentation remains challenging due to the limitations of its labeling information for unseen classes. Most previous approaches rely on extracting high-level feature maps from the frozen visual encoder to compute the pixel-wise similarity as a key prior guidance for the decoder. However, such a prior representation suffers from coarse granularity and poor generalization to new classes since these high-level feature maps have obvious category bias. In this work, we propose to replace the visual prior representation with the visual-text alignment capacity to capture more reliable guidance and enhance the model generalization. Specifically, we design two kinds of training-free prior information generation strategy that attempts to utilize the semantic alignment capability of the Contrastive Language-Image Pre-training model (CLIP) to locate the target class. Besides, to acquire more accurate prior guidance, we build a high-order relationship of attention maps and utilize it to refine the initial prior information. Experiments on both the PASCAL-5i and COCO-20i datasets show that our method obtains a clearly substantial improvement and reaches the new state-of-the-art performance. The code is available on the project website.

Open-vocabulary semantic segmentation (OVS) aims to segment images of arbitrary categories specified by class labels or captions. However, most previous best-performing methods, whether pixel grouping methods or region recognition methods, suffer from false matches between image features and category labels. We attribute this to the natural gap between the textual features and visual features. In this work, we rethink how to mitigate false matches from the perspective of image-to-image matching and propose a novel relation-aware intra-modal matching (RIM) framework for OVS based on visual foundation models. RIM achieves robust region classification by firstly constructing diverse image-modal reference features and then matching them with region features based on relation-aware ranking distribution. The proposed RIM enjoys several merits. First, the intra-modal reference features are better aligned, circumventing potential ambiguities that may arise in cross-modal matching. Second, the ranking-based matching process harnesses the structure information implicit in the inter-class relationships, making it more robust than comparing individually. Extensive experiments on three benchmarks demonstrate that RIM outperforms previous state-of-the-art methods by large margins, obtaining a lead of more than 10\% in mIoU on PASCAL VOC benchmark.

Saliency object ranking (SOR) has attracted significant attention recently. Previous methods usually failed to explicitly explore the saliency degree-related relationships between objects. In this paper, we propose a novel Domain Separation Graph Neural Network (DSGNN), which starts with separately extracting the shape and texture cues from each object, and builds an shape graph as well as a texture graph for all objects in the given image. Then, we propose a Shape-Texture Graph Domain Separation (STGDS) module to separate the task-relevant and irrelevant information of target objects by explicitly modelling the relationship between each pair of objects in terms of their shapes and textures, respectively. Furthermore, a Cross Image Graph Domain Separation (CIGDS) module is introduced to explore the saliency degree subspace that is robust to different scenes, aiming to create a unified representation for targets with the same saliency levels in different images. Importantly, our DSGNN automatically learns a multi-dimensional feature to represent each graph edge, allowing complex, diverse and ranking-related relationships to be modelled. Experimental results show that our DSGNN achieved the new state-of-the-art performance on both ASSR and IRSR datasets, with large improvements of 5.2\% and 4.1\% SA-SOR, respectively. Our code is provided in https://github.com/Wu-ZJ/DSGNN

Instance segmentation demands substantial labeling resources. This has prompted increased interest to explore the object discovery task as an unsupervised alternative. In particular, promising results were achieved in localizing instances using motion supervision only. However, the motion signal introduces complexities due to its inherent noise and sparsity, which constrains the effectiveness of current methodologies. In the present paper we propose DIOD (self DIstillation meets Object Discovery), the first method that places the motion-guided object discovery within a framework of continuous improvement through knowledge distillation, providing solutions to existing limitations (i) DIOD robustly eliminates the noise present in the exploited motion maps providing accurate motion-supervision (ii) DIOD leverages the discovered objects within an iterative pseudo-labeling framework, enriching the initial motion-supervision with static objects, which results in a cost-efficient increase in performance. Through experiments on synthetic and real-world datasets, we demonstrate the benefits of bridging the gap between object discovery and distillation, by significantly improving the state-of-the-art. This enhancement is also sustained across other demanding metrics so far reserved for supervised tasks. Code available upon acceptance.

Instance segmentation is data-hungry, and as model capacity increases, data scale becomes crucial for improving the accuracy. Most instance segmentation datasets today require costly manual annotation, limiting their data scale. Models trained on such data are prone to overfitting on the training set, especially for those rare categories. While recent works have delved into exploiting generative models to create synthetic datasets for data augmentation, these approaches do not efficiently harness the full potential of generative models.To address these issues, we introduce a more efficient strategy to construct generative datasets for data augmentation, termed DiverGen. Firstly, we provide an explanation of the role of generative data from the perspective of distribution discrepancy. We investigate the impact of different data on the distribution learned by the model. We argue that generative data can expand the data distribution that the model can learn, thus mitigating overfitting. Additionally, we find that the diversity of generative data is crucial for improving model performance and enhance it through various strategies, including category diversity, prompt diversity, and generative model diversity. With these strategies, we can scale the data to millions while maintaining the trend of model performance improvement. On the LVIS dataset, DiverGen significantly outperforms the strong model X-Paste, achieving +1.1 box AP and +1.1 mask AP across all categories, and +1.9 box AP and +2.5 mask AP for rare categories. Our codes are available at https://github.com/aim-uofa/DiverGen.

This paper revisits few-shot 3D point cloud semantic segmentation (FS-PCS), with a focus on two significant issues in the state-of-the-art: foreground leakage and sparse point distribution. The former arises from non-uniform point sampling, allowing models to distinguish the density disparities between foreground and background for easier segmentation. The latter results from sampling only 2,048 points, limiting semantic information and deviating from the real-world practice. To address these issues, we introduce a standardized FS-PCS setting, upon which a new benchmark is built. Moreover, we propose a novel FS-PCS model. While previous methods are based on feature optimization by mainly refining support features to enhance prototypes, our method is based on correlation optimization, referred to as Correlation Optimization Segmentation (COSeg). Specifically, we compute Class-specific Multi-prototypical Correlation (CMC) for each query point, representing its correlations to category prototypes. Then, we propose the Hyper Correlation Augmentation (HCA) module to enhance CMC. Furthermore, tackling the inherent property of few-shot training to incur base susceptibility for models, we propose to learn non-parametric prototypes for the base classes during training. The learned base prototypes are used to calibrate correlations for the background class through a Base Prototypes Calibration (BPC) module. Experiments on popular datasets demonstrate the superiority of COSeg over existing methods. The code is available at github.com/ZhaochongAn/COSeg.

We present S$^4$Former, a novel approach to training Vision Transformers for Semi-Supervised Semantic Segmentation (S$^4$). At its core, S$^4$Former employs a Vision Transformer within a classic teacher-student framework, and then leverages three novel technical ingredients: PatchShuffle as a parameter-free perturbation technique, Patch-Adaptive Self-Attention (PASA) as a fine-grained feature modulation method, and the innovative Negative Class Ranking (NCR) regularization loss. Based on these regularization modules aligned with Transformer-specific characteristics across the image input, feature, and output dimensions, S$^4$Former exploits the Transformer’s ability to capture and differentiate consistent global contextual information in unlabeled images. Overall, S$^4$Former not only defines a new state of the art in S$^4$ but also maintains a streamlined and scalable architecture. Being readily compatible with existing frameworks, S$^4$Former achieves strong improvements (up to 4.9\%) on benchmarks like Pascal VOC 2012, COCO, and Cityscapes, with varying numbers of labeled data. The code will be made publicly available.

We introduce Open3DIS a novel solution designed to tackle the problem of Open-Vocabulary Instance Segmentation within 3D scenes. Objects within 3D environments exhibit diverse shapes, scales, and colors, making precise instance-level identification a challenging task. Recent advancements in Open-Vocabulary scene understanding have made significant strides in this area by employing class-agnostic 3D instance proposal networks for object localization and learning queryable features for each 3D mask. While these methods produce high-quality instance proposals, they struggle with identifying small-scale and geometrically ambiguous objects. The key idea of our method is a new module that aggregates 2D instance masks across frames and maps them to geometrically coherent point cloud regions as high-quality object proposals addressing the above limitations. These are then combined with 3D class-agnostic instance proposals to include a wide range of objects in the real world. To validate our approach, we conducted experiments on three prominent datasets, including ScanNet200, S3DIS, and Replica, demonstrating significant performance gains in segmenting objects with diverse categories over the state-of-the-art approaches.

From an enormous amount of image-text pairs, large-scale vision-language models (VLMs) learn to implicitly associate image regions with words, which is vital for tasks such as image captioning and visual question answering. However, leveraging such pre-trained models for open-vocabulary semantic segmentation remains a challenge.In this paper, we propose a simple, yet extremely effective, training-free technique, Plug-and-Play Open-Vocabulary Semantic Segmentation (PnP-OVSS) for this task. PnP-OVSS leverages a VLM with direct text-to-image cross-attention and an image-text matching loss to produce semantic segmentation. However, cross-attention alone tends to over-segment, whereas cross-attention plus GradCAM tend to under-segment. To alleviate this issue, we introduce Salience Dropout; by iteratively dropping patches that the model is most attentive to, we are able to better resolve the entire extent of the segmentation mask. Compared to existing techniques, the proposed method does not require any neural network training and performs hyperparameter tuning without the need for any segmentation annotations, even for a validation set. PnP-OVSS demonstrates substantial improvements over a comparable baseline (+29.4\% on Pascal VOC, +13.2\% on Pascal Context, +14.0\% mIoU on MS COCO, +2.4\% on COCO Stuff) and even outperforms most baselines that conduct additional network training on top of pretrained VLMs.

Deep functional maps have emerged in recent years as a prominent learning-based framework for non-rigid shape matching problems. While early methods in this domain only focused on learning in the functional domain, the latest techniques have demonstrated that by promoting consistency between functional and pointwise maps leads to significant improvements in accuracy. Unfortunately, existing approaches rely heavily on the computation of large dense matrices arising from soft pointwise maps, which compromises their efficiency and scalability. To address this limitation, we introduce a novel memory-scalable and efficient functional map learning pipeline. By leveraging the specific structure of functional maps, we offer the possibility to achieve identical results without ever storing the pointwise map in memory. Furthermore, based on the same approach, we present a differentiable map refinement layer adapted from an existing axiomatic refinement algorithm. Unlike many functional map learning methods, which use this algorithm at a post-processing step, ours can be easily used at train time, enabling to enforce consistency between the refined and initial versions of the map. Our resulting approach is both simpler, more efficient and more numerically stable, by avoiding differentiation through a linear system, while achieving close to state-of-the-art results in challenging scenarios.

In recent interactive segmentation algorithms, previous probability maps are used as network input to help predictions in the current segmentation round. However, despite the utilization of previous masks, useful information contained in the probability maps is not well propagated to the current predictions. In this paper, to overcome this limitation, we propose a novel and effective framework for click-based interactive image segmentation, called MFP, to make full use of probability maps. We first modulate previous probability maps to enhance their representations of user-specified objects. Then, we feed the modulated probability maps as additional input to the segmentation network. We implement the proposed MFP algorithm based on the ResNet-34, HRNet-18, and ViT-B backbones and assess the performance extensively on four datasets: GrabCut, Berkeley, DAVIS, and SBD. It is demonstrated that MFP meaningfully outperforms the existing algorithms using identical backbones.

Coarse-to-fine 3D instance segmentation methods show weak performances compared to recent Grouping-based, Kernel-based and Transformer-based methods. We argue that this is due to two limitations: 1) Instance size overestimation by axis-aligned bounding box(AABB) 2) False negative error accumulation from inaccurate box to the refinement phase. In this work, we introduce $\textbf{Spherical Mask}$, a novel coarse-to-fine approach based on spherical representation, overcoming those two limitations with several benefits. Specifically, our coarse detection estimates each instance with a 3D polygon using a center and radial distance predictions, which avoids excessive size estimation of AABB. To cut the error propagation in the existing coarse-to-fine approaches, we virtually migrate points based on the polygon, allowing all foreground points, including false negatives, to be refined. During inference, the proposal and point migration modules run in parallel and are assembled to form binary masks of instances. We also introduce two margin-based losses for the point migration to enforce corrections for the false positives/negatives and cohesion of foreground points, significantly improving the performance. Experimental results from three datasets, such as ScanNetV2, S3DIS, and STPLS3D, show that our proposed method outperforms existing works, demonstrating the effectiveness of the new instance representation with spherical coordinates.

Consistency learning is a central strategy to tackle unlabeled data in semi-supervised medical image segmentation (SSMIS), which enforces the model to produce consistent predictions under the perturbation. However, most current approaches solely focus on utilizing a specific single perturbation, which can only cope with limited cases, while employing multiple perturbations simultaneously is hard to guarantee the quality of consistency learning. In this paper, we propose an Adaptive Bidirectional Displacement (ABD) approach to solve the above challenge. Specifically, we first design a bidirectional patch displacement based on reliable prediction confidence for unlabeled data to generate new samples, which can effectively suppress uncontrollable regions and still retain the influence of input perturbations. Meanwhile, to enforce the model to learn the potentially uncontrollable content, a bidirectional displacement operation with inverse confidence is proposed for the labeled images, which generates samples with more unreliable information to facilitate model learning. Extensive experiments show that ABD achieves new state-of-the-art performances for SSMIS, significantly improving different baselines. Source code is available at https://github.com/chy-upc/ABD.

Segment Anything Model (SAM) has emerged as a transformative approach in image segmentation, acclaimed for its robust zero-shot segmentation capabilities and flexible prompting system. Nonetheless, its performance is challenged by images with degraded quality. Addressing this limitation, we propose the Robust Segment Anything Model (RobustSAM), which enhances SAM's performance on low-quality images while preserving its promptability and zero-shot generalization. Our method leverages the pre-trained SAM model with only marginal parameter increments and computational requirements. The additional parameters of RobustSAM can be optimized within 30 hours on eight GPUs, demonstrating its feasibility and practicality for typical research laboratories. We also introduce the Robust-Seg dataset, a collection of 688K image-mask pairs with different degradations designed to train and evaluate our model optimally. Extensive experiments across various segmentation tasks and datasets confirm RobustSAM's superior performance, especially under zero-shot conditions, underscoring its potential for extensive real-world application. Additionally, our method has been shown to effectively improve the performance of SAM-based downstream tasks such as single image dehazing and deblurring. Code and dataset will be made available.

Camouflaged vision perception is an important vision task with numerous practical applications. Due to the expensive collection and labeling costs, this community struggles with a major bottleneck that the species category of its datasets is limited to a small number of object species. However, the existing camouflaged generation methods require specifying the background manually, thus failing to extend the camouflaged sample diversity in a low-cost manner. In this paper, we propose a Latent Background Knowledge Retrieval-Augmented Diffusion (LAKE-RED) for camouflaged image generation. To our knowledge, our contributions mainly include: (1) For the first time, we propose a camouflaged generation paradigm that does not need to receive any background inputs. (2) Our LAKE-RED is the first knowledge retrieval-augmented method with interpretability for camouflaged generation, in which we propose an idea that knowledge retrieval and reasoning enhancement are separated explicitly, to alleviate the task-specific challenges. Moreover, our method is not restricted to specific foreground targets or backgrounds, offering a potential for extending camouflaged vision perception to more diverse domains. (3) Experimental results demonstrate that our method outperforms the existing approaches, generating more realistic camouflage images.

Recent works utilize CLIP to perform the challenging unsupervised semantic segmentation task where only images without annotations are available. However, we observe that when adopting CLIP to such a pixel-level understanding task, unexpected bias occurs. Previous works didn't explicitly model such bias, which largely constrains the segmentation performance.In this paper, we propose to explicitly model and rectify the bias existing in CLIP to facilitate the unsupervised semantic segmentation. Specifically, we design learnable "Reference" prompt to represent class-preference bias and project positional embedding of vision transformer to represent space-preference bias. Via a simple element-wise subtraction, we rectify the logits of CLIP classifier.Based on the rectified logits, we generate a segmentation mask via a Gumbel-Softmax operation.Then a contrastive loss between masked visual feature and the text features of different classes is imposed to facilitate the effective bias modeling.To further improve the segmentation, we distill the knowledge from the rectified CLIP to the advanced segmentation architecture.Extensive experiments on standard benchmarks demonstrate that our method performs favorably against previous state-of-the-arts.

Open-vocabulary semantic segmentation presents the challenge of labeling each pixels within an image based on wide range of text descriptions. In this work, we introduce a novel cost-based approach to adapt vision-language foundation models, notably CLIP, for the intricate task of semantic segmentation. Through aggregating the cosine similarity score, i.e. the cost volume between image and text embeddings, our method potently adapts CLIP for segmenting seen and unseen classes by fine-tuning its encoders, addressing the challenges faced by existing methods in handling unseen classes. Building upon this, we explore methods to effectively aggregate the cost volume considering its multi-modal nature of being established between image and text embeddings. Furthermore, we examine various methods for efficiently fine-tuning CLIP. Our framework, dubbed CAT-Seg, shows state-of-the-art performance on standard benchmarks with significant margins, and further exerts strengths in more challenging scenarios from various domains.

Referring image segmentation (RIS) aims to segment the target referent described by natural language. Recently, large-scale pre-trained models, e.g., CLIP and SAM, have been successfully applied in many downstream tasks, but they are not well adapted to RIS task due to inter-task differences. In this paper, we propose a new prompt-driven framework named Prompt-RIS, which bridges CLIP and SAM end-to-end and transfers their rich knowledge and powerful capabilities to RIS task through prompt learning. To adapt CLIP to pixel-level task, we first propose a Cross-Modal Prompting method, which acquires more comprehensive vision-language interaction and fine-grained text-to-pixel alignment by performing bidirectional prompting. Then, the prompt-tuned CLIP generates masks, points, and text prompts for SAM to generate more accurate mask predictions. Moreover, we further propose Instance Contrastive Learning to improve the model's discriminability to different instances and robustness to diverse languages describing the same instance. Extensive experiments demonstrate that the performance of our method outperforms the state-of-the-art methods consistently in both general and open-vocabulary settings.

Image segmentation algorithms can be understood as a collection of pixel classifiers, for which the outcomes of nearby pixels are correlated. Classifier models can be calibrated using Inductive Conformal Prediction, but this requires holding back a sufficiently large calibration dataset for computing the distribution of non-conformity scores of the model's predictions. If one only requires only marginal calibration on the image level, this calibration set consists of all individual pixels in the images available for calibration. However, if the goal is to attain proper calibration for each individual pixel classifier, the calibration set consists of individual images. In a scenario where data are scarce (such as the medical domain), it may not always be possible to set aside sufficiently many images for this pixel-level calibration.The method we propose, dubbed "Kandinsky calibration", makes use of the spatial structure present in the distribution of natural images to simultaneously calibrate the classifiers of "similar" pixels. This can be seen as an intermediate approach between marginal (imagewise) and conditional (pixelwise) calibration, where non-conformity scores are aggregated over similar image regions, thereby making more efficient use of the images available for calibration. We run experiments on segmentation algorithms trained and calibrated on subsets of the public MS-COCO and Medical Decathlon datasets, demonstrating that Kandinsky calibration method can significantly improve the coverage. When compared to both pixelwise and imagewise calibration on little data, the Kandinsky method achieves much lower coverage errors, indicating the data efficiency of the Kandinsky calibration.

In the realm of food computing, segmenting ingredients from images poses substantial challenges due to the large intra-class variance among the same ingredients, the emergence of new ingredients, and the high annotation costs associated with large food segmentation datasets.Existing approaches primarily utilize a closed-vocabulary and static text embeddings setting. These methods often fall short in effectively handling the ingredients, particularly new and diverse ones. In response to these limitations, we introduce OVFoodSeg, a framework that adopts an open-vocabulary setting and enhances text embeddings with visual context.By integrating vision-language models (VLMs), our approach enriches text embedding with image-specific information through two innovative modules, \eg, an image-to-text learner FoodLearner and an Image-Informed Text Encoder.The training process of OVFoodSeg is divided into two stages: the pre-training of FoodLearner and the subsequent learning phase for segmentation. The pre-training phase equips FoodLearner with the capability to align visual information with corresponding textual representations that are specifically related to food, while the second phase adapts both the FoodLearner and the Image-Informed Text Encoder for the segmentation task.By addressing the deficiencies of previous models, OVFoodSeg demonstrates a significant improvement, achieving an 4.9\% increase in mean Intersection over Union (mIoU) on the FoodSeg103 dataset, setting a new milestone for food image segmentation.

With the immense growth of dataset sizes and computing resources in recent years, so-called foundation models have become popular in NLP and vision tasks. In this work, we propose to explore foundation models for the task of keypoint detection on 3D shapes. A unique characteristic of keypoint detection is that it requires semantic and geometric awareness while demanding high localization accuracy.To address this problem, we propose, first, to back-project features from large pre-trained 2D vision models onto 3D shapes and employ them for this task. We show that we obtain robust 3D features that contain rich semantic information and analyze multiple candidate features stemming from different 2D foundation models. Second, we employ a keypoint candidate optimization module which aims to match the average observed distribution of keypoints on the shape and is guided by the back-projected features. The resulting approach achieves a new state of the art for few-shot keypoint detection on the KeyPointNet dataset, almost doubling the performance of the previous best methods.

We present an approach for analyzing grouping information contained within a neural network's activations, permitting extraction of spatial layout and semantic segmentation from the behavior of large pre-trained vision models. Unlike prior work, our method conducts a wholistic analysis of a network's activation state, leveraging features from all layers and obviating the need to guess which part of the model contains relevant information. Motivated by classic spectral clustering, we formulate this analysis in terms of an optimization objective involving a set of affinity matrices, each formed by comparing features within a different layer. Solving this optimization problem using gradient descent allows our technique to scale from single images to dataset-level analysis, including, in the latter, both intra- and inter-image relationships. Analyzing a pre-trained generative transformer provides insight into the computational strategy learned by such models. Equating affinity with key-query similarity across attention layers yields eigenvectors encoding scene spatial layout, whereas defining affinity by value vector similarity yields eigenvectors encoding object identity. This result suggests that key and query vectors coordinate attentional information flow according to spatial proximity (a 'where' pathway), while value vectors refine a semantic category representation (a 'what' pathway).

We study the underexplored but fundamental vision problem of machine understanding of abstract freehand scene sketches. We introduce a sketch encoder that results in semantically-aware feature space, which we evaluate by testing its performance on a semantic sketch segmentation task. To train our model we rely only on the availability of bitmap sketches with their brief captions and do not require any pixel-level annotations. To obtain generalization to a large set of sketches and categories, we build on a vision transformer encoder pretrained with the CLIP model. We freeze the text encoder and perform visual-prompt tuning of the visual encoder branch while introducing a set of critical modifications. Firstly, we augment the classical key-query (k-q) self-attention blocks with value-value (v-v) self-attention blocks. Central to our model is a two-level hierarchical network design that enables efficient semantic disentanglement: The first level ensures holistic scene sketch encoding, and the second level focuses on individual categories. We, then, in the second level of the hierarchy, introduce a cross-attention between textual and visual branches. Our method outperforms zero-shot CLIP pixel accuracy of segmentation results by 37 points, reaching an accuracy of 85.5\% on the FS-COCO sketch dataset. Finally, we conduct a user study that allows us to identify further improvements needed over our method to reconcile machine and human understanding of scene sketches.

The open-vocabulary image segmentation task involves partitioning images into semantically meaningful segments and classifying them with flexible text-defined categories. The recent vision-based foundation models such as the Segment Anything Model (SAM) have shown superior performance in generating class-agnostic image segments. The main challenge in open-vocabulary image segmentation now lies in accurately classifying these segments into text-defined categories. In this paper, we introduce the Universal Segment Embedding (USE) framework to address this challenge. This framework is comprised of two key components: 1) a data pipeline designed to efficiently curate a large amount of segment-text pairs at various granularities, and 2) a universal segment embedding model that enables precise segment classification into a vast range of text-defined categories. The USE model can not only help open-vocabulary image segmentation but also facilitate other downstream tasks (e.g., querying and ranking). Through comprehensive experimental studies on semantic segmentation and part segmentation benchmarks, we demonstrate that the USE framework outperforms state-of-the-art open-vocabulary segmentation methods.

Diffusion models trained on large-scale datasets have achieved remarkable progress in image synthesis. However, due to the randomness in the diffusion process, they often struggle with handling diverse low-level tasks that require details preservation. To overcome this limitation, we present a new Diff-Plugin framework to enable a single pre-trained diffusion model to generate high-fidelity results across a variety of low-level tasks. Specifically, we first propose a lightweight Task-Plugin module with a dual branch design to provide task-specific priors, guiding the diffusion process in preserving image content.We then propose a Plugin-Selector that can automatically select different Task-Plugins based on the text instruction, allowing users to edit images by indicating multiple low-level tasks with natural language.We conduct extensive experiments on 8 low-level vision tasks. The results demonstrate the superiority of Diff-Plugin over existing methods, particularly in real-world scenarios. Our ablations further validate that Diff-Plugin is stable, schedulable, and supports robust training across different dataset sizes.

We present $\mathcal{X}^3$ (pronounced XCube), a novel generative model for high-resolution sparse 3D voxel grids with arbitrary attributes. Our model can generate millions of voxels with a finest effective resolution of up to $1024^3$ in a feed-forward fashion without time-consuming test-time optimization. To achieve this, we employ a hierarchical voxel latent diffusion model which generates progressively higher resolution grids in a coarse-to-fine manner using a custom framework built on the highly efficient VDB data structure. Apart from generating high-resolution objects, we demonstrate the effectiveness of XCube on large outdoor scenes at scales of 100m$\times$100m with a voxel size as small as 10cm. We observe clear qualitative and quantitative improvements over past approaches. In addition to unconditional generation, we show that our model can be used to solve a variety of tasks such as user-guided editing, scene completion from a single scan, and text-to-3D.

Novel view synthesis for dynamic scenes is still a challenging problem in computer vision and graphics. Recently, Gaussian splatting has emerged as a robust technique to represent static scenes and enable high-quality and real-time novel view synthesis. Building upon this technique, we propose a new representation that explicitly decomposes the motion and appearance of dynamic scenes into sparse control points and dense Gaussians, respectively. Our key idea is to use sparse control points, significantly fewer in number than the Gaussians, to learn compact 6 DoF transformation bases, which can be locally interpolated through learned interpolation weights to yield the motion field of 3D Gaussians. We employ a deformation MLP to predict time-varying 6 DoF transformations for each control point, which reduces learning complexities, enhances learning abilities, and facilitates obtaining temporal and spatial coherent motion patterns. Then, we jointly learn the 3D Gaussians, the canonical space locations of control points, and the deformation MLP to reconstruct the appearance, geometry, and dynamics of 3D scenes. During learning, the location and number of control points are adaptively adjusted to accommodate varying motion complexities in different regions, and an ARAP loss following the principle of as rigid as possible is developed to enforce spatial continuity and local rigidity of learned motions. Finally, thanks to the explicit sparse motion representation and its decomposition from appearance, our method can enable user-controlled motion editing while retaining high-fidelity appearances. Extensive experiments demonstrate that our approach outperforms existing approaches on novel view synthesis with a high rendering speed and enables novel appearance-preserved motion editing applications.

We describe a novel method, StyLitGAN, for relighting and resurfacing images in the absence of labeled data. StyLitGAN generates images with realistic lighting effects, including cast shadows, soft shadows, inter-reflections, and glossy effects, without the need for paired or CGI data. StyLitGAN uses an intrinsic image method to decompose an image, followed by a search of the latent space of a pretrained StyleGAN to identify a set of directions. By prompting the model to fix one component (e.g., albedo) and vary another (e.g., shading), we generate relighted images by adding the identified directions to the latent style codes. Quantitative metrics of change in albedo and lighting diversity allow us to choose effective directions using a forward selection process. Qualitative evaluation confirms the effectiveness of our method.

State-of-the-art man-made shape generative models usually adopt established generative models under a suitable implicit shape representation. A common theme is to perform distribution alignment, which does not explicitly model important shape priors. As a result, many synthetic shapes are not connected. Other synthetic shapes present problems of physical stability and geometric feasibility. This paper introduces a novel latent diffusion shape-generative model guided by a quality check that outputs a score of a latent code. The scoring function employs a learned function that provides a geometric feasibility score and a deterministic procedure to quantify a physical stability score. The key to our approach is a new diffusion procedure that combines the discrete empirical data distribution and a continuous distribution induced by the quality checker. We introduce a principled approach to determine the tradeoff parameters for learning the denoising network at different noise levels. We also present an efficient strategy that avoids evaluating the score for each synthetic shape during the optimization procedure. Experimental results show that our approach outperforms state-of-the-art shape generations quantitatively and qualitatively on ShapeNet-v2.

We present Image Sculpting, a new framework for editing 2D images by incorporating tools from 3D geometry and graphics. This approach differs markedly from existing methods, which are confined to 2D spaces and typically rely on textual instructions, leading to ambiguity and limited control. Image Sculpting converts 2D objects into 3D, enabling direct interaction with their 3D geometry. Post-editing, these objects are re-rendered into 2D, merging into the original image to produce high-fidelity results through a coarse-to-fine enhancement process. The framework supports precise, quantifiable, and physically-plausible editing options such as pose editing, rotation, translation, 3D composition, carving, and serial addition. It marks an initial step towards combining the creative freedom of generative models with the precision of graphics pipelines.

This paper presents Paint3D, a novel coarse-to-fine generative framework that is capable of producing high-resolution, lighting-less, and diverse 2K UV texture maps for untextured 3D meshes conditioned on text or image inputs. The key challenge addressed is generating high-quality textures without embedded illumination information, which allows the textures to be re-lighted or re-edited within modern graphics pipelines. To achieve this, our method first leverages a pre-trained depth-aware 2D diffusion model to generate view-conditional images and perform multi-view texture fusion, producing an initial coarse texture map. However, as 2D models cannot fully represent 3D shapes and disable lighting effects, the coarse texture map exhibits incomplete areas and illumination artifacts. To resolve this, we train separate UV Inpainting and UVHD diffusion models specialized for the shape-aware refinement of incomplete areas and the removal of illumination artifacts. Through this coarse-to-fine process, Paint3D can produce high-quality 2K UV textures that maintain semantic consistency while being lighting-less, significantly advancing the state-of-the-art in texturing 3D objects.

Content-aware graphic layout generation aims to automatically arrange visual elements along with a given content, such as an e-commerce product image. In this paper, we argue that the current layout generation approaches suffer from the limited training data for the high-dimensional layout structure. We show that a simple retrieval augmentation can significantly improve the generation quality. Our model, which is named Retrieval-Augmented Layout Transformer (RALF), retrieves nearest neighbor layout examples based on an input image and feeds these results into an autoregressive generator. Our model can apply retrieval augmentation to various controllable generation tasks and yield high-quality layouts within a unified architecture. Our extensive experiments show that RALF successfully generates content-aware layouts in both constrained and unconstrained settings and significantly outperforms the baselines.

At the core of portrait photography is the search for ideal lighting and viewpoint. The process often requires advanced knowledge in photography and an elaborate studio setup. In this work, we propose Holo-Relighting, a volumetric relighting method that is capable of synthesizing novel viewpoints, and novel lighting from a single image. Holo-Relighting leverages the pretrained 3D GAN (EG3D) to reconstruct geometry and appearance from an input portrait as a set of 3D-aware features. We design a relighting module conditioned on a given lighting to process these features, and predict a relit 3D representation in the form of a tri-plane, which can render to an arbitrary viewpoint through volume rendering. Besides viewpoint and lighting control, Holo-Relighting also takes the head pose as a condition to enable head-pose-dependent lighting effects. With these novel designs, Holo-Relighting can generate complex non-Lambertian lighting effects (e.g., specular highlights and cast shadows) without using any explicit physical lighting priors. We train Holo-Relighting with data captured with a light stage, and propose two data-rendering techniques to improve the data quality for training the volumetric relighting system. Through quantitative and qualitative experiments, we demonstrate Holo-Relighting can achieve state-of-the-arts relighting quality with better photorealism, 3D consistency and controllability.

Neural fields have emerged as a powerful and broadly applicable method for representing signals. However, in contrast to classical discrete digital signal processing, the portfolio of tools to process such representations is still severely limited and restricted to Euclidean domains. In this paper, we address this problem by showing how a probabilistic re-interpretation of neural fields can enable their training and inference processes to become "filter-aware". The formulation we propose not only merges training and filtering in an efficient way, but also generalizes beyond the familiar Euclidean coordinate spaces to the more general set of smooth manifolds and convolutions induced by the actions of Lie groups. We demonstrate how this framework can enable novel integrations of signal processing techniques for neural field applications on both Euclidean domains, such as images and audio, as well as non-Euclidean domains, such as rotations and rays. A noteworthy benefit of our method is its applicability. Our method can be summarized as primarily a modification of the loss function, and in most cases does not require changes to the network architecture or the inference process.

Decomposing an object's appearance into representations of its materials and the surrounding illumination is difficult, even when the object's 3D shape is known beforehand. This problem is especially challenging for diffuse objects: it is ill-conditioned because diffuse materials severely blur incoming light, and it is ill-posed because diffuse materials under high-frequency lighting can be indistinguishable from shiny materials under low-frequency lighting. We show that it is possible to recover precise materials and illumination---even from diffuse objects---by exploiting unintended shadows, like the ones cast onto an object by the photographer who moves around it. These shadows are a nuisance in most previous inverse rendering pipelines, but here we exploit them as signals that improve conditioning and help resolve material-lighting ambiguities. We present a method based on differentiable Monte Carlo ray tracing that uses images of an object to jointly recover its spatially-varying materials, the surrounding illumination environment, and the shapes of the unseen light occluders who inadvertently cast shadows upon it.

This paper focuses on synthesizing high-quality and complete textures directly on the surface of 3D models within 3D space. 2D diffusion-based methods face challenges in generating 2D texture maps due to the infinite possibilities of UV mapping for a given 3D mesh. Utilizing point clouds helps circumvent variations arising from diverse mesh topologies and UV mappings. Nevertheless, achieving dense point clouds to accurately represent texture details poses a challenge due to limited computational resources. To address these challenges, we propose an efficient octree-based diffusion pipeline called TexOct. Our method starts by sampling a point cloud from the surface of a given 3D model, with each point containing texture noise values. We utilize an octree structure to efficiently represent this point cloud. Additionally, we introduce an innovative octree-based diffusion model that leverages the denoising capabilities of the Denoising Diffusion Probabilistic Model (DDPM). This model gradually reduces the texture noise on the octree nodes, resulting in the restoration of fine texture. Experimental results on ShapeNet demonstrate that TexOct effectively generates high-quality 3D textures in both unconditional and text / image-conditional scenarios.

Micro-mesh ($\mu$-mesh) is a new graphics primitive for compact representation of extreme geometry, consisting of a low-polygon base mesh enriched by per micro-vertex displacement. A new generation of GPUs supports this structure with hardware evolution on $\mu$-mesh ray tracing, achieving real-time rendering in pixel level geometric details. In this article, we present a differentiable framework to convert standard meshes into this efficient format, offering a holistic scheme in contrast to the previous stage-based methods. In our construction context, a $\mu$-mesh is defined where each base triangle is a parametric primitive, which is then reparameterized with Laplacian operators for efficient geometry optimization. Our framework offers numerous advantages for high-quality $\mu$-mesh production: (1) end-to-end geometry optimization and displacement baking; (2) enabling the differentiation of renderings with respect to $\mu$-mesh for faithful reprojectability; (3) high scalability for integrating useful features for $\mu$-mesh production and rendering, such as minimizing shell volume, maintaining the isotropy of the base mesh, and visual-guided adaptive level of detail. Extensive experiments on $\mu$-mesh construction for a large set of high-resolution meshes demonstrate the superior quality achieved by the proposed scheme.

We present TextureDreamer, a novel image-guided texture synthesis method to transfer relightable textures from a small number of input images (3 to 5) to target 3D shapes across arbitrary categories. Texture creation is a pivotal challenge in vision and graphics. Industrial companies hire experienced artists to manually craft textures for 3D assets.Classical methods require densely sampled views and accurately aligned geometry, while learning-based methods are confined to category-specific shapes within the dataset. In contrast, TextureDreamer can transfer highly detailed, intricate textures from real-world environments to arbitrary objects with only a few casually captured images, potentially significantly democratizing texture creation. Our core idea, $\textit{personalized geometry-aware score distillation (PGSD)}$, draws inspiration from recent advancements in diffuse models, including personalized modeling for texture information extraction, variational score distillation for detailed appearance synthesis, and explicit geometry guidance with ControlNet.Our integration and several essential modifications substantially improve the texture quality. Experiments on real images spanning different categories show that TextureDreamer can successfully transfer highly realistic, semantic meaningful texture to arbitrary objects, surpassing the visual quality of previous state-of-the-art.

We present As-Plausible-as-Possible (APAP) mesh deformation technique that leverages 2D diffusion priors to preserve the plausibility of a mesh under user-controlled deformation. Our framework uses per-face Jacobians to represent mesh deformations, where mesh vertex coordinates are computed via a differentiable Poisson Solve. The deformed mesh is rendered, and the resulting 2D image is used in the Score Distillation Sampling (SDS) process, which enables extracting meaningful plausibility priors from a pretrained 2D diffusion model. To better preserve the identity of the edited mesh, we fine-tune our 2D diffusion model with LoRA. Gradients extracted by SDS and a user-prescribed handle displacement are then backpropagated to the per-face Jacobians, and we use iterative gradient descent to compute the final deformation that balances between the user edit and the output plausibility. We evaluate our method with 2D and 3D meshes and demonstrate qualitative and quantitative improvements when using plausibility priors over geometry-preservation or distortion-minimization priors used by previous techniques. Our project page is at: https://as-plausible-aspossible.github.io/

A sketch is one of the most intuitive and versatile tools humans use to convey their ideas visually. An animated sketch opens another dimension to the expression of ideas and is widely used by designers for a variety of purposes.Animating sketches is a laborious process, requiring extensive experience and professional design skills.In this work, we present a method that automatically adds motion to a single-subject sketch (hence, ``breathing life into it''), merely by providing a text prompt indicating the desired motion.The output is a short animation provided in vector representation, which can be easily edited.Our method does not require extensive training, but instead leverages the motion prior of a large pretrained text-to-video diffusion model using a score-distillation loss to guide the placement of strokes. To promote natural and smooth motion and to better preserve the sketch's appearance, we model the learned motion through two components. The first governs small local deformations and the second controls global affine transformations.Surprisingly, we find that even models that struggle to generate sketch videos on their own can still serve as a useful backbone for animating abstract representations.

We propose a neural reflectance model (NeuBRDF) that offers highly versatile material representation, yet with light memory and neural computation consumption towards achieving real-time rendering. The results depicted in Fig. 1, rendered at full HD resolution on a contemporary desktop machine, demonstrate that our system achieves real-time performance with a wide variety of appearances, which is approached by the following two designs. Firstly, recognizing that the bidirectional reflectance is distributed in a sparse high-dimensional space, we propose to project the BRDF into two low-dimensional components, i.e. two hemisphere feature-grids for incoming and outgoing directions, respectively. Secondly, we distribute learnable neural reflectance primitives on our highly-tailored spherical surface grid. These primitives offer informative features for each hemisphere component and reduce the complexity of the feature learning network, leading to fast evaluation. These primitives are centrally stored in a codebook and can be shared across multiple grids and even across materials, based on low-cost indices stored in material-specific spherical surface grids. Our NeuBRDF, agnostic to the material, provides a unified framework for representing a variety of materials consistently. Comprehensive experimental results on measured BRDF compression, Monte Carlo simulated BRDF acceleration, and extension to spatially varying effects demonstrate the superior quality and generalizability achieved by the proposed scheme.

We present Paint-it, a text-driven high-fidelity texture map synthesis method for 3D meshes via neural re-parameterized texture optimization. Paint-it synthesizes texture maps from a text description by synthesis-through-optimization, exploiting the Score-Distillation Sampling (SDS). We observe that directly applying SDS yields undesirable texture quality due to its noisy gradients. We reveal the importance of texture parameterization when using SDS. Specifically, we propose Deep Convolutional Physically-Based Rendering (DC-PBR) parameterization, which re-parameterizes the physically-based rendering (PBR) texture maps with randomly initialized convolution-based neural kernels, instead of a standard pixel-based parameterization. We show that DC-PBR inherently schedules the optimization curriculum according to texture frequency and naturally filters out the noisy signals from SDS. In experiments, Paint-it obtains remarkable quality PBR texture maps within 15 min., given only a text description. We demonstrate the generalizability and practicality of Paint-it by synthesizing high-quality texture maps for large-scale mesh datasets and showing test-time applications such as relighting and material control using a popular graphics engine. Code will be publicly available.

It is time-consuming to render high-resolution images in applications such as video games and virtual reality, and thus super-resolution technologies become increasingly popular for real-time rendering. However, it is challenging to preserve sharp texture details, keep the temporal stability and avoid the ghosting artifacts in real-time super-resolution rendering. To address this issue, we introduce radiance demodulation to separate the rendered image or radiance into a lighting component and a material component, considering the fact that the light component is smoother than the rendered image so that the high-resolution material component with detailed textures can be easily obtained. We perform the super-resolution on the lighting component only and re-modulate it with the high-resolution material component to obtain the final super-resolution image with more texture details. A reliable warping module is proposed by explicitly masking the occluded regions to avoid the ghosting artifacts. To further enhance the temporal stability, we design a frame-recurrent neural network and a temporal loss to aggregate the previous and current frames, which can better capture the spatial-temporal consistency among reconstructed frames. As a result, our method is able to produce temporally stable results in real-time rendering with high-quality details, even in the challenging 4 × 4 super-resolution scenarios.

The realism of digital avatars is crucial in enabling telepresence applications with self-expression and customization. A key aspect of this realism originates from the physical accuracy of both a true-to-life body shape and clothing.While physical simulations can produce high-quality, realistic motions for clothed humans, they require precise estimation of body shape and high-quality garment assets with associated physical parameters for cloth simulations. However, manually creating these assets and calibrating their parameters is labor-intensive and requires specialized expertise. To address this gap, we propose DiffAvatar, a novel approach that performs body and garment co-optimization using differentiable simulation. By integrating physical simulation into the optimization loop and accounting for the complex non-linear behavior of cloth and its intricate interaction with the body, our framework recovers body and garment geometry and extracts important material parameters in a physically plausible way. Our experiments demonstrate that our approach generates realistic clothing and body shape that can be easily used in downstream applications.

In this paper, we propose a method to extract Physically-Based-Rendering (PBR) materials from a single real-world image. We do so in two steps: first, we map regions of the image to material concepts using a diffusion model, which allows sampling of texture images resembling each material in the scene. Second, we benefit from a separate network to decompose the generated textures into Spatially Varying BRDFs (SVBRDFs), providing us with materials ready to be used in rendering applications. Our approach builds on existing synthetic material libraries with SVBRDF ground truth, but also exploits a diffusion-generated RGB texture dataset to allow generalization to new samples using unsupervised domain adaptation (UDA). Our contributions are thoroughly evaluated on synthetic and real-world datasets. We further demonstrate the applicability of our method for editing 3D scenes with materials estimated from real photographs. The code and models will be made open-source.

We introduce PhysGaussian, a new method that seamlessly integrates physically grounded Newtonian dynamics within 3D Gaussians to achieve high-quality novel motion synthesis. Employing a customized Material Point Method (MPM), our approach enriches 3D Gaussian kernels with physically meaningful kinematic deformation and mechanical stress attributes, all evolved in line with continuum mechanics principles. A defining characteristic of our method is the seamless integration between physical simulation and visual rendering: both components utilize the same 3D Gaussian kernels as their discrete representations. This negates the necessity for triangle/tetrahedron meshing, marching cubes, cage meshes, or any other geometry embedding, highlighting the principle of "what you see is what you simulate (WS^2)". Our method demonstrates exceptional versatility across a wide variety of materials--including elastic entities, plastic metals, non-Newtonian fluids, and granular materials--showcasing its strong capabilities in creating diverse visual content with novel viewpoints and movements.

We present differentiable point-based inverse rendering, DPIR, an analysis-by-synthesis method that processes images captured under diverse illuminations to estimate shape and spatially-varying BRDF. To this end, we adopt point-based rendering, eliminating the need for multiple samplings per ray, typical of volumetric rendering, thus significantly enhancing the speed of inverse rendering. To realize this idea, we devise a hybrid point-volumetric representation for geometry and a regularized basis-BRDF representation for reflectance.The hybrid geometric representation enables fast rendering through point-based splatting while retaining the geometric details and stability inherent to SDF-based representations. The regularized basis-BRDF mitigates the ill-posedness of inverse rendering stemming from limited light-view angular samples. We also propose an efficient shadow detection method using point-based shadow map rendering.Our extensive evaluations demonstrate that DPIR outperforms prior works in terms of reconstruction accuracy, computational efficiency, and memory footprint. Furthermore, our explicit point-based representation and rendering enables intuitive geometry and reflectance editing. The code will be publicly available.

We develop a theory for the representation of opaque solids as volumes. Starting from a stochastic representation of opaque solids as random indicator functions, we prove the conditions under which such solids can be modeled using exponential volumetric transport. We also derive expressions for the volumetric attenuation coefficient as a functional of the probability distributions of the underlying indicator functions. We generalize our theory to account for isotropic and anisotropic scattering at different parts of the solid, and for representations of opaque solids as stochastic implicit surfaces. We derive our volumetric representation from first principles, which ensures that it satisfies physical constraints such as reciprocity and reversibility. We use our theory to explain, compare, and correct previous volumetric representations, as well as propose meaningful extensions that lead to improved performance in 3D reconstruction tasks.

Progress in lighting estimation is tracked by computing existing image quality assessment (IQA) metrics on images from standard datasets. While this may appear to be a reasonable approach, we demonstrate that doing so does not correlate to human preference when the estimated lighting is used to relight a virtual scene into a real photograph. To study this, we design a controlled psychophysical experiment where human observers must choose their preference amongst rendered scenes lit using a set of lighting estimation algorithms selected from the recent literature, and use it to analyse how these algorithms perform according to human perception. Then, we demonstrate that none of the most popular IQA metrics from the literature, taken individually, correctly represent human perception. Finally, we show that by learning a combination of existing IQA metrics, we can more accurately represent human preference. This provides a new perceptual framework to help evaluate future lighting estimation algorithms.

Intelligent generation of vector graphics has very promising applications in the fields of advertising and logo design, artistic painting, animation production, etc. However, current mainstream vector image generation methods based on diffusion models lack the encoding of image appearance information that is associated with the original vector representation and therefore lose valid supervision signal from the strong correlation between the discrete vector parameter (drawing instruction) sequence and the target shape/structure of the corresponding pixel image. On the one hand, the generation process based on pure vector domain completely ignores the similarity measurement between shape parameter (and their combination) and the paired pixel image appearance pattern; on the other hand, two-stage methods (i.e., generation-and-vectorization) based on pixel diffusion followed by differentiable image-to-vector translation suffer from wrong error-correction signal caused by approximate gradients. To address the above issues, we propose a novel generation framework based on dual-domain (vector-pixel) diffusion with cross-modality impulse signals from each other. First, in each diffusion step, the current representation extracted from the other domain is used as a condition variable to constrain the subsequent sampling operation, yielding shape-aware new parameterizations; second, independent supervision signals from both domains avoid the gradient error accumulation problem caused by cross-domain representation conversion. Extensive experimental results on popular benchmarks including font and icon datasets demonstrate the great advantages of our proposed framework in terms of generated shape quality.

Creating high-quality materials in computer graphics is a challenging and time-consuming task, which requires great expertise. To simplify this process, we introduce MatFuse, a unified approach that harnesses the generative power of diffusion models for creation and editing of 3D materials.Our method integrates multiple sources of conditioning, including color palettes, sketches, text, and pictures, enhancing creative possibilities and granting fine-grained control over material synthesis.Additionally, MatFuse enables map-level material editing capabilities through latent manipulation by means of a multi-encoder compression model which learns a disentangled latent representation for each map.We demonstrate the effectiveness of MatFuse under multiple conditioning settings and explore the potential of material editing. Finally, we assess the quality of the generated materials both quantitatively in terms of CLIP-IQA and FID scores and qualitatively by conducting a user study.Source code for training MatFuse and supplemental materials are publicly available at https://gvecchio.com/matfuse.

We present a simple yet effective technique to estimate lighting in a single input image. Current techniques rely heavily on HDR panorama datasets to train neural networks to regress an input with limited field-of-view to a full environment map. However, these approaches often struggle with real-world, uncontrolled settings due to the limited diversity and size of their datasets. To address this problem, we leverage diffusion models trained on billions of standard images to render a chrome ball into the input image. Despite its simplicity, this task remains challenging: the diffusion models often insert incorrect or inconsistent objects and cannot readily generate images in HDR format. Our research uncovers a surprising relationship between the appearance of chrome balls and the initial diffusion noise map, which we utilize to consistently generate high-quality chrome balls. We further fine-tune an LDR diffusion model (Stable Diffusion XL) with LoRA, making it able to perform exposure bracketing for HDR light estimation. Our method produces convincing light estimates across diverse settings and demonstrates superior generalization to in-the-wild scenarios.

We introduce TexTile, a novel differentiable metric to quantify the degree upon which a texture image can be concatenated with itself without introducing repeating artifacts (i.e., the tileability). Existing methods for tileable texture synthesis focus on general texture quality, but lack explicit analysis of the intrinsic repeatability properties of a texture. In contrast, our TexTile metric effectively evaluates the tileable properties of a texture, opening the door to more informed synthesis and analysis of tileable textures. Under the hood, TexTile is formulated as a binary classifier carefully built from a large dataset of textures of different styles, semantics, regularities, and human annotations.Key to our method is a set of architectural modifications to baseline pre-train image classifiers to overcome their shortcomings at measuring tileability, along with a custom data augmentation and training regime aimed at increasing robustness and accuracy. We demonstrate that TexTile can be plugged into different state-of-the-art texture synthesis methods, including diffusion-based strategies, and generate tileable textures while keeping or even improving the overall texture quality. Furthermore, we show that TexTile can objectively evaluate any tileable texture synthesis method, whereas the current mix of existing metrics produces uncorrelated scores which heavily hinders progress in the field.

We show that physics-based simulations can be seamlessly integrated with NeRF to generate high-quality elastodynamics of real-world objects. Unlike existing methods, we discretize nonlinear hyperelasticity in a meshless way, obviating the necessity for intermediate auxiliary shape proxies like a tetrahedral mesh or voxel grid. A quadratic generalized moving least square (Q-GMLS) is employed to capture nonlinear dynamics and large deformation on the implicit model. Such meshless integration enables versatile simulations of complex and codimensional shapes. We adaptively place the least-square kernels according to the NeRF density field to significantly reduce the complexity of the nonlinear simulation. As a result, physically realistic animations can be conveniently synthesized using our method for a wide range of hyperelastic materials at an interactive rate.

In this paper, we address the problem of efficient point searching and sampling for volume neural rendering. Within this realm, two typical approaches are employed: rasterization and ray tracing. The rasterization-based methods enable real-time rendering at the cost of increased memory and lower fidelity. In contrast, the ray-tracing-based methods yield superior quality but demand longer rendering time. We solve this problem by our HashPoint method combining these two strategies, leveraging rasterization for efficient point searching and sampling, and ray marching for rendering. Our method optimizes point searching by rasterizing points within the camera's view, organizing them in a hash table, and facilitating rapid searches. Notably, we accelerate the rendering process by adaptive sampling on the primary surface encountered by the ray. Our approach yields substantial speed-up for a range of state-of-the-art ray-tracing-based methods, maintaining equivalent or superior accuracy across synthetic and real test datasets. The code will be available at https://jiahao-ma.github.io/hashpoint/.

We present 3D Paintbrush, a technique for automatically texturing local semantic regions on meshes via text descriptions. Our method is designed to operate directly on meshes, producing texture maps which seamlessly integrate into standard graphics pipelines. We opt to simultaneously produce a localization map (to specify the edit region) and a texture map which conforms to it. This approach improves the quality of both the localization and the stylization. To enhance the details and resolution of the textured area, we leverage multiple stages of a cascaded diffusion model to supervise our local editing technique with generative priors learned from images at different resolutions. Our technique, referred to as Cascaded Score Distillation (CSD), simultaneously distills scores at multiple resolutions in a cascaded fashion, enabling control over both the granularity and global understanding of the supervision. We demonstrate the effectiveness of 3D Paintbrush to locally texture different semantic regions on a variety of shapes.

In recent years, there has been a growing interest in training Neural Networks to approximate Unsigned Distance Fields (UDFs) for representing open surfaces in the context of 3D reconstruction. However, UDFs are non-differentiable at the zero level set which leads to significant errors in distances and gradients, generally resulting in fragmented and discontinuous surfaces. In this paper, we propose to learn a hyperbolic scaling of the unsigned distance field, which defines a new Eikonal problem with distinct boundary conditions. This allows our formulation to integrate seamlessly with state-of-the-art continuously differentiable implicit neural representation networks, largely applied in the literature to represent signed distance fields. Our approach not only addresses the challenge of open surface representation but also demonstrates significant improvement in reconstruction quality and training performance. Moreover, the unlocked field's differentiability allows the accurate computation of essential topological properties such as normal directions and curvatures, pervasive in downstream tasks such as rendering. Through extensive experiments, we validate our approach across various data sets and against competitive baselines. The results demonstrate enhanced accuracy and up to an order of magnitude increase in speed compared to previous methods.

We present Diff3F as a simple, robust, and class-agnostic feature descriptor that can be computed for untextured input shapes (meshes or point clouds). Our method distills diffusion features from image foundational models onto input shapes. Specifically, we use the input shapes to produce depth and normal maps as guidance for conditional image synthesis, and in the process produce (diffusion) features in 2D that we subsequently lift and aggregate on the original surface. Our key observation is that even if the conditional image generations obtained from multi-view rendering of the input shapes are inconsistent, the associated image features are robust and can be directly aggregated across views. This produces semantic features on the input shapes, without requiring additional data or training. We perform extensive experiments on multiple benchmarks (SHREC'19, SHREC'20, and TOSCA) and demonstrate that our features, being semantic instead of geometric, produce reliable correspondence across both isometeric and non-isometrically related shape families.

Recent advances in 3D face stylization have made significant strides in few to zero-shot settings. However, the degree of stylization achieved by existing methods is often not sufficient for practical applications because they are mostly based on statistical 3D Morphable Models (3DMM) with limited variations. To this end, we propose a method that can produce a highly stylized 3D face model with desired topology. Our methods train a surface deformation network with 3DMM and translate its domain to the target style using a paired exemplar.The network achieves stylization of the 3D face mesh by mimicking the style of the target using a differentiable renderer and directional CLIP losses. Additionally, during the inference process, we utilize a Mesh Agnostic Encoder (MAGE) that takes deformation target, a mesh of diverse topologies as input to the stylization process and encodes its shape into our latent space.The resulting stylized face model can be animated by commonly used 3DMM blend shapes.A set of quantitative and qualitative evaluations demonstrate that our method can produce highly stylized face meshes according to a given style and output them in a desired topology. We also demonstrate example applications of our method including image-based stylized avatar generation, linear interpolation of geometric styles, and facial animation of stylized avatars.

Bokeh is widely used in photography to draw attention to the subject while effectively isolating distractions in the background. Computational methods can simulate bokeh effects without relying on a physical camera lens, but the inaccurate lens modeling in existing filtering-based methods leads to artifacts that need post-processing or learning-based methods to fix. We propose Dr.Bokeh, a novel rendering method that addresses the issue by directly correcting the defect that violates the physics in the current filtering-based bokeh rendering equation. Dr.Bokeh first preprocesses the input RGBD to obtain a layered scene representation. Dr.Bokeh then takes the layered representation and user-defined lens parameters to render photo-realistic lens blur based on the novel occlusion-aware bokeh rendering method. Experiments show that the non-learning based renderer Dr.Bokeh outperforms state-of-the-art bokeh rendering algorithms in terms of photo-realism.In addition, extensive quantitative and qualitative evaluations show the more accurate lens model further pushes the limit of a closely related field depth-from-defocus.

We present DiffInDScene, a novel framework for tackling the problem of high-quality 3D indoor scene generation, which is challenging due to the complexity and diversity of the indoor scene geometry.Although diffusion-based generative models have previously demonstrated impressive performance in image generation and object-level 3D generation, they have not yet been applied to room-level 3D generation due to their computationally intensive costs.In DiffInDScene, we propose a cascaded 3D diffusion pipeline that is efficient and possesses strong generative performance for Truncated Signed Distance Function (TSDF). The whole pipeline is designed to run on a sparse occupancy space in a coarse-to-fine fashion.Inspired by KinectFusion's incremental alignment and fusion of local TSDF volumes, we propose a diffusion-based SDF fusion approach that iteratively diffuses and fuses local TSDF volumes, facilitating the generation of an entire room environment. The generated results demonstrate that our work is capable to achieve high-quality room generation directly in three-dimensional space, starting from scratch.In addition to the scene generation, the final part of DiffInDScene can be used as a post-processing module to refine the 3D reconstruction results from multi-view stereo.According to the user study, the mesh quality generated by our DiffInDScene can even outperform the ground truth mesh provided by ScanNet.

We present a lightweight solution for estimating spatially-coherent indoor lighting from a single RGB image. Previous methods for estimating illumination using volumetric representations have overlooked the sparse distribution of light sources in space, necessitating substantial memory and computational resources for achieving high-quality results. We introduce a unified, voxel octree-based illumination estimation framework to produce 3D spatially-coherent lighting. Additionally, a differentiable voxel octree cone tracing rendering layer is proposed to eliminate regular volumetric representation throughout the entire process and ensure the retention of features across different frequency domains.This reduction significantly decreases spatial usage and required floating-point operations without substantially compromising precision.Experimental results demonstrate that our approach achieves high-quality coherent estimation with minimal cost compared to previous methods.

Recently, text-guided scalable vector graphics (SVGs) synthesis has shown promise in domains such as iconography and sketch. However, existing text-to-SVG generation methods lack editability and struggle with visual quality and result diversity. To address these limitations, we propose a novel text-guided vector graphics synthesis method called SVGDreamer.SVGDreamer incorporates a semantic-driven image vectorization (SIVE) process that enables the decomposition of synthesis into foreground objects and background, thereby enhancing editability. Specifically, the SIVE process introduces attention-based primitive control and an attention-mask loss function for effective control and manipulation of individual elements.Additionally, we propose a Vectorized Particle-based Score Distillation (VPSD) approach to address issues of shape over-smoothing, color over-saturation, limited diversity, and slow convergence of the existing text-to-SVG generation methods by modeling SVGs as distributions of control points and colors. Furthermore, VPSD leverages a reward model to re-weight vector particles, which improves aesthetic appeal and accelerates convergence.Extensive experiments are conducted to validate the effectiveness of SVGDreamer, demonstrating its superiority over baseline methods in terms of editability, visual quality, and diversity.Project page: \href{https://ximinng.github.io/SVGDreamer-project/}{https://ximinng.github.io/SVGDreamer-project/}

Recent years have witnessed considerable achievements in editing images with text instructions. When applying these editors to dynamic scene editing, the new-style scene tends to be temporally inconsistent due to the frame-by-frame nature of these 2D editors. To tackle this issue, we propose Control4D, a novel approach for high-fidelity and temporally consistent 4D portrait editing. Control4D is built upon an efficient 4D representation with a 2D diffusion-based editor. Instead of using direct supervision from the editor, our method learns a 4D generator from it and avoids the inconsistent supervision signals. Specifically, we employ a discriminator to learn the generation distribution based on the edited images and then update the generator with the discrimination signals. For more stable training, multi-level information is extracted from the edited images and used to facilitate the learning of the generator. Experimental results show that Control4D surpasses previous approaches and achieves more photo-realistic and consistent 4D editing performances.

Recent text-to-3D methods employing diffusion models have made significant advancements in 3D human generation. However, these approaches face challenges due to the limitations of text-to-image diffusion models, which lack an understanding of 3D structures. Consequently, these methods struggle to achieve high-quality human generation, resulting in smooth geometry and cartoon-like appearances. In this paper, we propose HumanNorm, a novel approach for high-quality and realistic 3D human generation. The main idea is to enhance the model's 2D perception of 3D geometry by learning a normal-adapted diffusion model and a normal-aligned diffusion model. The normal-adapted diffusion model can generate high-fidelity normal maps corresponding to user prompts with view-dependent and body-aware text. The normal-aligned diffusion model learns to generate color images aligned with the normal maps, thereby transforming physical geometry details into realistic appearance. Leveraging the proposed normal diffusion model, we devise a progressive geometry generation strategy and a multi-step Score Distillation Sampling (SDS) loss to enhance the performance of 3D human generation. Comprehensive experiments substantiate HumanNorm’s ability to generate 3D humans with intricate geometry and realistic appearances. HumanNorm outperforms existing text-to-3D methods in both geometry and texture quality.

Creating high-quality and interactive virtual environments, such as games and simulators, often involves complex and costly manual modeling processes. In this paper, we present Video2Game, a novel approach that automatically converts videos of real-world scenes into realistic and interactive game environments. At the heart of our system are three core components: (i) a neural radiance fields (NeRF) module that effectively captures the geometry and visual appearance of the scene; (ii) a mesh module that distills the knowledge from NeRF for faster rendering; and (iii) a physics module that models the interactions and physical dynamics among the objects. By following the carefully designed pipeline, one can construct an interactable and actionable digital replica of the real world. We benchmark our system on both indoor and large-scale outdoor scenes. We show that we can not only produce highly-realistic renderings in real-time, but also build interactive games on top.

The success of denoising diffusion models in representing rich data distributions over 2D raster images has prompted research on extending them to other data representations, such as vector graphics. Unfortunately due to their variable structure and scarcity of vector training data, directly applying diffusion models on this domain remains a challenging problem. Using workarounds like optimization via Score Distillation Sampling (SDS) is also fraught with difficulty, as vector representations are non-trivial to directly optimize and tend to result in implausible geometries such as redundant or self-intersecting shapes. NIVeL addresses these challenges by reinterpreting the problem on an alternative, intermediate domain which preserves the desirable properties of vector graphics – mainly sparsity of representation and resolution-independence. This alternative domain is based on neural implicit fields expressed in a set of decomposable, editable layers. Based on our experiments, NIVeL produces text-to-vector graphics results of significantly better quality than the state-of-the-art.

Existing NeRF-based inverse rendering methods suppose that scenes are exclusively illuminated by distant light sources, neglecting the potential influence of emissive sources within a scene. In this work, we confront this limitation using LDR multi-view images captured with emissive sources turned on and off. Two key issues must be addressed: 1) ambiguity arising from the limited dynamic range along with unknown lighting details, and 2) the expensive computational cost in volume rendering to backtrace the paths leading to final object colors. We present a novel approach, ESR-NeRF, leveraging neural networks as learnable functions to represent ray-traced fields. By training networks to satisfy light transport segments, we regulate outgoing radiances, progressively identifying emissive sources while being aware of reflection areas. The results on scenes encompassing emissive sources with various properties demonstrate the superiority of ESR-NeRF in qualitative and quantitative ways. Our approach also extends its applicability to the scenes devoid of emissive sources, achieving lower CD metrics on the DTU dataset.

Recent methods such as Score Distillation Sampling (SDS) and Variational Score Distillation (VSD) using 2D diffusion models for text-to-3D generation have demonstrated impressive generation quality. However, the long generation time of such algorithms significantly degrades the user experience. To tackle this problem, we propose DreamPropeller, a drop-in acceleration algorithm that can be wrapped around any existing text-to-3D generation pipeline based on score distillation. Our framework generalizes Picard iterations, a classical algorithm for parallel sampling an ODE path, and can account for non-ODE paths such as momentum-based gradient updates and changes in dimensions during the optimization process as in many cases of 3D generation. We show that our algorithm trades parallel compute for wallclock time and empirically achieves up to 4.7x speedup with a negligible drop in generation quality for all tested frameworks.

We present GenesisTex, a novel method for synthesizing textures for 3D geometries from text descriptions. GenesisTex adapts the pretrained image diffusion model to texture space by texture space sampling. Specifically, we maintain a latent texture map for each viewpoint, which is updated with predicted noise on the rendering of the corresponding viewpoint. The sampled latent texture maps are then decoded into a final texture map. During the sampling process, we focus on both global and local consistency across multiple viewpoints: global consistency is achieved through the integration of style consistency mechanisms within the noise prediction network, and low-level consistency is achieved by dynamically aligning latent textures. Finally, we apply reference-based inpainting and img2img on denser views for texture refinement. Our approach overcomes the limitations of slow optimization in distillation-based methods and instability in inpainting-based methods. Experiments on meshes from various sources demonstrate that our method surpasses the baseline methods quantitatively and qualitatively.

Current diffusion or flow-based generative models for 3D shapes divide to two: distilling pre-trained 2D image diffusion models, and training directly on 3D shapes. When training a diffusion or flow models on 3D shapes a crucial design choice is the shape representation. An effective shape representation needs to adhere three design principles: it should allow an efficient conversion of large 3D datasets to the representation form; it should provide a good tradeoff of approximation power versus number of parameters; and it should have a simple tensorial form that is compatible with existing powerful neural architectures. While standard 3D shape representations such as volumetric grids and point clouds do not adhere to all these principles simultaneously, we advocate in this paper a new representation that does. We introduce Mosaic-SDF (M-SDF): a simple 3D shape representation that approximates the Signed Distance Function (SDF) of a given shape by using a set of local grids spread near the shape's boundary. The M-SDF representation is fast to compute for each shape individually making it readily parallelizable; it is parameter efficient as it only covers the space around the shape's boundary; and it has a simple matrix form, compatible with Transformer-based architectures. We demonstrate the efficacy of the M-SDF representation by using it to train a 3D generative flow model including class-conditioned generation with the ShapeNetCore-V2 (3D Warehouse) dataset, and text-to-3D generation using a dataset of about 600k caption-shape pairs.

A Neural Radiance Field (NeRF) encodes the specific relation of 3D geometry and appearance of a scene. We here ask the question whether we can transfer the appearance from a source NeRF onto a target 3D geometry in a semantically meaningful way, such that the resulting new NeRF retains the target geometry but has an appearance that is an analogy to the source NeRF. To this end, we generalize classic image analogies from 2D images to NeRFs. We leverage correspondence transfer along semantic affinity that is driven by semantic features from large, pre-trained 2D image models to achieve multi-view consistent appearance transfer. Our method allows exploring the mix-and-match product space of 3D geometry and appearance. We show that our method outperforms traditional stylization-based methods and that a large majority of users prefer our method over several typical baselines.

Mesh denoising (MD) is a critical task in geometry processing, as meshes from scanning or AIGC techniques are susceptible to noise contamination. The challenge of MD lies in the diverse nature of mesh facets in terms of geometric characteristics and noise distributions.Despite recent advancements in deep learning-based MD methods, existing MD networks typically neglect the consideration of geometric characteristics and noise distributions. In this paper, we propose Hyper-MD, a hyper-network-based approach that addresses this limitation by dynamically customizing denoising parameters for each facet based on its noise intensity and geometric characteristics. Specifically, Hyper-MD is composed of a hyper-network and an MD network. For each noisy facet, the hyper-network takes two angles as input to customize parameters for the MD network. These two angles are specially defined to reveal the noise intensity and geometric characteristics of the current facet, respectively. The MD network receives a facet patch as input, and outputs the denoised normal using the customized parameters. Experimental results on synthetic and real-scanned meshes demonstrate that Hyper-MD outperforms state-of-the-art mesh denoising methods.

Despite exciting progress in automatic 3D reconstruction from images, excessive and irregular triangular faces in the resulting meshes still constitute a significant challenge when it comes to adoption in practical artist workflows. Therefore, we propose a method to extract regular quad-dominant meshes from posed images. More specifically, we generate a high-quality 3D model through decomposition into an easily editable quad-dominant mesh with pixel-level details such as displacement, materials, and lighting. To enable end-to-end learning of shape and quad topology, we QUADify a neural implicit representation using our novel differentiable re-meshing objective. Distinct from previous work, our method exploits artifact-free Catmull-Clark subdivision combined with vertex displacement to extract pixel-level details linked to the base geometry. Finally, we apply differentiable rendering techniques for material and lighting decomposition to optimize for image reconstruction. Our experiments show the benefits of end-to-end re-meshing and that our method yields state-of-the-art geometric accuracy while providing lightweight meshes with displacements and textures that are directly compatible with professional renderers and game engines.

This paper introduces SfmCAD, a novel unsupervised network that learns the Sketch-based Feature Modeling operations used in modern CAD workflow to reconstruct 3D shapes. Given a 3D shape represented as voxels, SfmCAD learns a neural typed sketch+path representation, including 2D sketches of feature primitives and their 3D sweeping paths without supervision, for inferring feature-based CAD programs. This approach bridges the gap between detail-oriented shape reconstruction and the simplicity and control intrinsic to primitive extraction. By utilizing 2D sketches to represent local shape details and sweeping paths to encapsulate the structure of the shape, SfmCAD achieves an interpretable decoupling of shape structure and local details. By manipulating the parametric 2D sketch and 3D path, SfmCAD facilitates users in making distinct modifications to both the geometric and structural characteristics of the shape. We demonstrate the effectiveness of our method by applying SfmCAD to many different types of objects, such as CAD parts, ShapeNet objects, and tree shapes. Extensive comparisons show that SfmCAD produces compact and faithful 3D reconstructions with superior quality than existing alternatives. The code will be released to facilitate future research.

Learning-based isosurface extraction methods have recently emerged as a robust and efficient alternative to axiomatic techniques. However, the vast majority of such approaches rely on supervised training with axiomatically computed ground truths, thus potentially inheriting biases and data artefacts of the corresponding axiomatic methods. Steering away from such dependencies, we propose a self-supervised training scheme to the Neural Dual Contouring meshing framework, resulting in our method: Self-Supervised Dual Contouring (SDC). Instead of optimizing predicted mesh vertices with supervised training we use two novel self-supervised loss functions which encourage the consistency between distances to the generated mesh and the given input SDF, and align corresponding face normals. Meshes reconstructed by SDC not only exhibit an improved consistency with the input SDFs but also surpass existing data-driven methods in capturing intricate details, while being more robust to possible irregularities in the input. Furthermore, we use the same self-supervised training objective linking inferred mesh and input SDF, to regualize the training process of Deep Implicit Networks (DINs). We demonstrate that the resulting DINs produce higher quality implicit functions, ultimately leading to more accurate and detail-preserving surfaces compared to prior baselines for different input modalities. Finally, we demonstrate that our self-supervised losses improve meshing performance in the single-view reconstruction task by enabling joint training of predicted SDF and resulting output mesh.

Efficiently representing and reconstructing the 3D geometry of trees remains a challenging problem in computer vision and graphics. We propose a novel approach for generating realistic tree models from single-view photographs. We cast the 3D information inference problem to a semantic voxel diffusion process, which converts an input image of a tree to a novel Semantic Voxel Structure (SVS) in 3D space. The SVS encodes the geometric appearance and semantic structural information (e.g., classifying trunks, branches, and leaves), having the distinct advantage of retaining the intricate internal features of trees. Tailored to the SVS, we also present a new hybrid tree modeling approach by combining structure-oriented branch reconstruction and self-organization-based foliage reconstruction. We validate our approach by using images from both synthetic and real trees. The comparison results show that our approach can better preserve tree details and achieve more realistic and accurate reconstruction results than previous methods. We will open-source the code and data to facilitate future research.

We present HAAR, a new strand-based generative model for 3D human hairstyles. Specifically, based on textual inputs, HAAR produces 3D hairstyles that could be used as production-level assets in modern computer graphics engines. Current AI-based generative models take advantage of powerful 2D priors to reconstruct 3D content in the form of point clouds, meshes, or volumetric functions. However, by using the 2D priors, they are intrinsically limited to only recovering the visual parts. Highly occluded hair structures can not be reconstructed with those methods, and they only model the "outer shell", which is not ready to be used in physics-based rendering or simulation pipelines. In contrast, we propose a first text-guided generative method that uses 3D hair strands as an underlying representation. Leveraging 2D visual question-answering (VQA) systems, we automatically annotate synthetic hair models that are generated from a small set of artist-created hairstyles. This allows us to train a latent diffusion model that operates in a common hairstyle UV space. In qualitative and quantitative studies, we demonstrate the capabilities of the proposed model and compare it to existing hairstyle generation approaches. For results, please refer to our project page https://haar.is.tue.mpg.de/.

Reverse engineering in the realm of Computer-Aided Design (CAD) has been a longstanding aspiration, though not yet entirely realized. Its primary aim is to uncover the CAD process behind a physical object given its 3D scan. We propose CAD-SIGNet, an end-to-end trainable and auto-regressive architecture to recover the design history of a CAD model represented as a sequence of sketch-and-extrusion from an input point cloud. Our model learns visual-language representations by layer-wise cross-attention between point cloud and CAD language embedding. In particular, a new Sketch instance Guided Attention (SGA) module is proposed in order to reconstruct the fine-grained details of the sketches. Thanks to its auto-regressive nature, CAD-SIGNet not only reconstructs a unique full design history of the corresponding CAD model given an input point cloud but also provides multiple plausible design choices. This allows for an interactive reverse engineering scenario by providing designers with multiple next step choices along with the design process. Extensive experiments on publicly available CAD datasets showcase the effectiveness of our approach against existing baseline models in two settings, namely, full design history recovery and conditional auto-completion from point clouds.

We propose a new class of generative diffusion models, called functional diffusion. In contrast to previous work, functional diffusion works on samples that are represented by functions with a continuous domain. Functional diffusion can be seen as an extension of classical diffusion models to the infinite-dimensional domain. Functional diffusion is very versatile as images, videos, audio, 3D shapes, deformations, etc., can be handled by the same framework with minimal changes. In addition, functional diffusion is especially suited for irregular data or data defined in non-standard domains. In our work, we derive the necessary foundations for functional diffusion and propose a first implementation based on the transformer architecture. We show generative results on complicated signed distance functions and deformation functions defined on 3D shape surfaces.