English

What Makes Good Synthetic Training Data for Zero-Shot Stereo Matching?

Computer Vision and Pattern Recognition 2026-03-02 v3

Abstract

Synthetic datasets are a crucial ingredient for training stereo matching networks, but the question of what makes a stereo dataset effective remains underexplored. We investigate the design space of synthetic datasets by varying the parameters of a procedural dataset generator, and report the effects on zero-shot stereo matching performance using standard benchmarks. We validate our findings by collecting the best settings and creating a large-scale dataset. Training only on this dataset achieves better performance than training on a mixture of widely used datasets, and is competitive with training on the FoundationStereo dataset, with the additional benefit of open-source generation code and an accompanying parameter analysis to enable further research. We open-source our system at https://github.com/princeton-vl/InfinigenStereo to enable further research on procedural stereo datasets.

Keywords

Cite

@article{arxiv.2504.16930,
  title  = {What Makes Good Synthetic Training Data for Zero-Shot Stereo Matching?},
  author = {David Yan and Alexander Raistrick and Jia Deng},
  journal= {arXiv preprint arXiv:2504.16930},
  year   = {2026}
}

Comments

Accepted to CVPR 2026

R2 v1 2026-06-28T23:08:53.448Z