On the confidence of stereo matching in a deep-learning era: a quantitative evaluation

TL;DR

This paper provides a comprehensive review and evaluation of confidence estimation methods in stereo matching over ten years, comparing traditional and deep learning approaches across multiple datasets.

Contribution

It offers the first extensive comparison of confidence measures with deep stereo networks, highlighting their strengths and limitations.

Findings

Learning-based confidence measures have specific strengths and limitations.

Traditional confidence measures perform variably across different stereo algorithms.

Deep learning methods show promising results but also face challenges in reliability.

Abstract

Stereo matching is one of the most popular techniques to estimate dense depth maps by finding the disparity between matching pixels on two, synchronized and rectified images. Alongside with the development of more accurate algorithms, the research community focused on finding good strategies to estimate the reliability, i.e. the confidence, of estimated disparity maps. This information proves to be a powerful cue to naively find wrong matches as well as to improve the overall effectiveness of a variety of stereo algorithms according to different strategies. In this paper, we review more than ten years of developments in the field of confidence estimation for stereo matching. We extensively discuss and evaluate existing confidence measures and their variants, from hand-crafted ones to the most recent, state-of-the-art learning based methods. We study the different behaviors of each measure when applied to a pool of different stereo algorithms and, for the first time in literature, when paired with a state-of-the-art deep stereo network. Our experiments, carried out on five different standard datasets, provide a comprehensive overview of the field, highlighting in particular both strengths and limitations of learning-based strategies.