Domain Generalized Stereo Matching with Uncertainty-guided Data Augmentation

TL;DR

This paper introduces an uncertainty-guided data augmentation method for stereo matching that improves cross-domain generalization by perturbing image statistics and enforcing feature consistency, applicable to any network.

Contribution

We propose a novel uncertainty-guided data augmentation technique that models domain variability through perturbations of image statistics and enhances feature invariance in stereo matching.

Findings

Significantly improves cross-domain stereo matching performance

Effective across multiple challenging benchmarks

Architecture-agnostic and easy to integrate

Abstract

State-of-the-art stereo matching (SM) models trained on synthetic data often fail to generalize to real data domains due to domain differences, such as color, illumination, contrast, and texture. To address this challenge, we leverage data augmentation to expand the training domain, encouraging the model to acquire robust cross-domain feature representations instead of domain-dependent shortcuts. This paper proposes an uncertainty-guided data augmentation (UgDA) method, which argues that the image statistics in RGB space (mean and standard deviation) carry the domain characteristics. Thus, samples in unseen domains can be generated by properly perturbing these statistics. Furthermore, to simulate more potential domains, Gaussian distributions founded on batch-level statistics are poposed to model the unceratinty of perturbation direction and intensity. Additionally, we further enforce feature consistency between original and augmented data for the same scene, encouraging the model to learn structure aware, shortcuts-invariant feature representations. Our approach is simple, architecture-agnostic, and can be integrated into any SM networks. Extensive experiments on several challenging benchmarks have demonstrated that our method can significantly improve the generalization performance of existing SM networks.