Co-Teaching: An Ark to Unsupervised Stereo Matching

TL;DR

This paper introduces CoT-Stereo, an unsupervised stereo matching method that uses a co-teaching framework to improve accuracy near occlusions, outperforming existing methods on KITTI benchmarks.

Contribution

The paper presents a novel co-teaching framework for unsupervised stereo matching, enhancing robustness and accuracy near occlusions compared to prior approaches.

Findings

Outperforms state-of-the-art unsupervised methods on KITTI benchmarks.

Achieves higher accuracy and speed in stereo matching.

Effectively handles occlusions through interactive network teaching.

Abstract

Stereo matching is a key component of autonomous driving perception. Recent unsupervised stereo matching approaches have received adequate attention due to their advantage of not requiring disparity ground truth. These approaches, however, perform poorly near occlusions. To overcome this drawback, in this paper, we propose CoT-Stereo, a novel unsupervised stereo matching approach. Specifically, we adopt a co-teaching framework where two networks interactively teach each other about the occlusions in an unsupervised fashion, which greatly improves the robustness of unsupervised stereo matching. Extensive experiments on the KITTI Stereo benchmarks demonstrate the superior performance of CoT-Stereo over all other state-of-the-art unsupervised stereo matching approaches in terms of both accuracy and speed. Our project webpage is https://sites.google.com/view/cot-stereo.