DCVSMNet: Double Cost Volume Stereo Matching Network

TL;DR

DCVSMNet is a fast and accurate stereo matching network that uses dual cost volumes and a coupling module to improve geometry understanding, achieving competitive results with strong generalization.

Contribution

Introduces a novel dual cost volume architecture with a coupling module for enhanced stereo matching performance.

Findings

Achieves 67 ms inference time.

Outperforms CGI-Stereo and BGNet in accuracy.

Demonstrates strong generalization across datasets.

Abstract

We introduce Double Cost Volume Stereo Matching Network(DCVSMNet) which is a novel architecture characterised by by two small upper (group-wise) and lower (norm correlation) cost volumes. Each cost volume is processed separately, and a coupling module is proposed to fuse the geometry information extracted from the upper and lower cost volumes. DCVSMNet is a fast stereo matching network with a 67 ms inference time and strong generalization ability which can produce competitive results compared to state-of-the-art methods. The results on several bench mark datasets show that DCVSMNet achieves better accuracy than methods such as CGI-Stereo and BGNet at the cost of greater inference time.