Sparse LiDAR Assisted Self-supervised Stereo Disparity Estimation

TL;DR

This paper introduces a self-supervised stereo disparity estimation method that leverages sparse LiDAR data to improve efficiency and generalization, avoiding expensive 4D cost volumes.

Contribution

It proposes integrating sparse LiDAR points into iterative disparity updates and training in a self-supervised manner for better real-world applicability.

Findings

Achieves comparable results with state-of-the-art methods

Reduces computational burden by avoiding 4D cost volume

Enhances generalization through self-supervised training

Abstract

Deep stereo matching has made significant progress in recent years. However, state-of-the-art methods are based on expensive 4D cost volume, which limits their use in real-world applications. To address this issue, 3D correlation maps and iterative disparity updates have been proposed. Regarding that in real-world platforms, such as self-driving cars and robots, the Lidar is usually installed. Thus we further introduce the sparse Lidar point into the iterative updates, which alleviates the burden of network updating the disparity from zero states. Furthermore, we propose training the network in a self-supervised way so that it can be trained on any captured data for better generalization ability. Experiments and comparisons show that the presented method is effective and achieves comparable results with related methods.