Skip to yearly menu bar Skip to main content


Workshop

Pixel-level Video Understanding in the Wild Challenge

Jiaxu Miao · Yunchao Wei · Zongxin Yang · Yi Yang · Si Liu · Yi Zhu · Elisa Ricci · Cees Snoek

West 301

Sun 18 Jun, 1:30 p.m. PDT

Keywords:  Video understanding  

Pixel-level Scene Understanding is one of the fundamental problems in computer vision, which aims at recognizing object classes, masks and semantics of each pixel in the given image. Since the real-world is actually video-based rather than a static state, learning to perform video semantic/panoptic segmentation is more reasonable and practical for realistic applications. To advance the semantic/panoptic segmentation task from images to videos, we present two large-scale datasets (VSPW and VIPSeg) and a competition in this workshop, aiming at performing the challenging yet practical Pixel-level Video Understanding in the Wild (PVUW).

Chat is not available.