Improved Point Transformation Methods For Self-Supervised Depth Prediction

TL;DR

This paper introduces a z-buffering algorithm and a negative depth loss to improve unsupervised monocular depth prediction, addressing occlusion handling and erroneous shallow depths, leading to better performance on KITTI.

Contribution

It presents a novel z-buffering method for occlusion handling and a loss function for negative depths, enhancing existing depth prediction frameworks.

Findings

Z-buffering improves occlusion handling in depth estimation.

Negative depth loss reduces shallow depth prediction errors.

Method achieves better results on KITTI dataset.

Abstract

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.