Improved Point Transformation Methods For Self-Supervised Depth Prediction
This work addresses a specific limitation in unsupervised depth estimation for computer vision applications, representing an incremental improvement over existing methods.
The paper tackled the problem of handling occluded points in self-supervised depth prediction by introducing a z-buffering algorithm and a loss function to penalize negative depth predictions, resulting in improved performance on the KITTI dataset.
Given stereo or egomotion image pairs, a popular and successful method for unsupervised learning of monocular depth estimation is to measure the quality of image reconstructions resulting from the learned depth predictions. Continued research has improved the overall approach in recent years, yet the common framework still suffers from several important limitations, particularly when dealing with points occluded after transformation to a novel viewpoint. While prior work has addressed this problem heuristically, this paper introduces a z-buffering algorithm that correctly and efficiently handles occluded points. Because our algorithm is implemented with operators typical of machine learning libraries, it can be incorporated into any existing unsupervised depth learning framework with automatic support for differentiation. Additionally, because points having negative depth after transformation often signify erroneously shallow depth predictions, we introduce a loss function to penalize this undesirable behavior explicitly. Experimental results on the KITTI data set show that the z-buffer and negative depth loss both improve the performance of a state of the art depth-prediction network.