Lite Any Stereo: Efficient Zero-Shot Stereo Matching

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jing, Junpeng, Luo, Weixun, Mao, Ye, Mikolajczyk, Krystian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915938393128960
author Jing, Junpeng
Luo, Weixun
Mao, Ye
Mikolajczyk, Krystian
author_facet Jing, Junpeng
Luo, Weixun
Mao, Ye
Mikolajczyk, Krystian
contents Recent advances in stereo matching have focused on accuracy, often at the cost of significantly increased model size. Traditionally, the community has regarded efficient models as incapable of zero-shot ability due to their limited capacity. In this paper, we introduce Lite Any Stereo, a stereo depth estimation framework that achieves strong zero-shot generalization while remaining highly efficient. To this end, we design a compact yet expressive backbone to ensure scalability, along with a carefully crafted hybrid cost aggregation module. We further propose a three-stage training strategy on million-scale data to effectively bridge the sim-to-real gap. Together, these components demonstrate that an ultra-light model can deliver strong generalization, ranking 1st across four widely used real-world benchmarks. Remarkably, our model attains accuracy comparable to or exceeding state-of-the-art non-prior-based accurate methods while requiring less than 1% computational cost, setting a new standard for efficient stereo matching.
format Preprint
id arxiv_https___arxiv_org_abs_2511_16555
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Lite Any Stereo: Efficient Zero-Shot Stereo Matching
Jing, Junpeng
Luo, Weixun
Mao, Ye
Mikolajczyk, Krystian
Computer Vision and Pattern Recognition
Recent advances in stereo matching have focused on accuracy, often at the cost of significantly increased model size. Traditionally, the community has regarded efficient models as incapable of zero-shot ability due to their limited capacity. In this paper, we introduce Lite Any Stereo, a stereo depth estimation framework that achieves strong zero-shot generalization while remaining highly efficient. To this end, we design a compact yet expressive backbone to ensure scalability, along with a carefully crafted hybrid cost aggregation module. We further propose a three-stage training strategy on million-scale data to effectively bridge the sim-to-real gap. Together, these components demonstrate that an ultra-light model can deliver strong generalization, ranking 1st across four widely used real-world benchmarks. Remarkably, our model attains accuracy comparable to or exceeding state-of-the-art non-prior-based accurate methods while requiring less than 1% computational cost, setting a new standard for efficient stereo matching.
title Lite Any Stereo: Efficient Zero-Shot Stereo Matching
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.16555