Skip to yearly menu bar Skip to main content


Poster

Learning a Depth Covariance Function

Eric Dexheimer · Andrew J. Davison

West Building Exhibit Halls ABC 073

Abstract:

We propose learning a depth covariance function with applications to geometric vision tasks. Given RGB images as input, the covariance function can be flexibly used to define priors over depth functions, predictive distributions given observations, and methods for active point selection. We leverage these techniques for a selection of downstream tasks: depth completion, bundle adjustment, and monocular dense visual odometry.

Chat is not available.