Inferring Distributions Over Depth from a Single Image
This addresses the need for robust geometric scene understanding in autonomous vehicles by providing uncertainty-aware depth estimation, which is incremental over existing Bayesian methods.
The paper tackles the problem of depth estimation from a single image by recasting it as discrete binary classification to output un-normalized distributions over possible depths per pixel, enabling reliable and efficient capture of multi-modal depth distributions in ambiguous cases. Results show accurate depth predictions, significantly better uncertainty estimations than prior art while running near real-time, and improvements in 3D map reconstruction by reducing artifacts and enhancing accuracy and memory efficiency.
When building a geometric scene understanding system for autonomous vehicles, it is crucial to know when the system might fail. Most contemporary approaches cast the problem as depth regression, whose output is a depth value for each pixel. Such approaches cannot diagnose when failures might occur. One attractive alternative is a deep Bayesian network, which captures uncertainty in both model parameters and ambiguous sensor measurements. However, estimating uncertainties is often slow and the distributions are often limited to be uni-modal. In this paper, we recast the continuous problem of depth regression as discrete binary classification, whose output is an un-normalized distribution over possible depths for each pixel. Such output allows one to reliably and efficiently capture multi-modal depth distributions in ambiguous cases, such as depth discontinuities and reflective surfaces. Results on standard benchmarks show that our method produces accurate depth predictions and significantly better uncertainty estimations than prior art while running near real-time. Finally, by making use of uncertainties of the predicted distribution, we significantly reduce streak-like artifacts and improves accuracy as well as memory efficiency in 3D map reconstruction.