Stereo-Matching Knowledge Distilled Monocular Depth Estimation Filtered by Multiple Disparity Consistency
This addresses a key bottleneck in self-supervised depth estimation for applications like robotics and autonomous driving, though it is incremental as it builds on existing distillation methods.
The paper tackles the problem of error propagation in self-supervised monocular depth estimation by proposing a method to filter errors in pseudo-depth maps using multiple disparity consistency, eliminating the need for ground truth or training. Experimental results show it outperforms previous methods, particularly in challenging areas like textureless regions and occlusion boundaries.
In stereo-matching knowledge distillation methods of the self-supervised monocular depth estimation, the stereo-matching network's knowledge is distilled into a monocular depth network through pseudo-depth maps. In these methods, the learning-based stereo-confidence network is generally utilized to identify errors in the pseudo-depth maps to prevent transferring the errors. However, the learning-based stereo-confidence networks should be trained with ground truth (GT), which is not feasible in a self-supervised setting. In this paper, we propose a method to identify and filter errors in the pseudo-depth map using multiple disparity maps by checking their consistency without the need for GT and a training process. Experimental results show that the proposed method outperforms the previous methods and works well on various configurations by filtering out erroneous areas where the stereo-matching is vulnerable, especially such as textureless regions, occlusion boundaries, and reflective surfaces.