Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Fast-FoundationStereo: Real-Time Zero-Shot Stereo Matching
Bowen Wen, Shaurya Dewan, Stan Birchfield
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| distill hybrid FoundationStereo backbone priors into student backbone (ablation) | Fast-FoundationStereo · Middlebury-H | BP-2: 2.87 -> 2.20 | Improved |
| distill backbone priors (ablation) | Fast-FoundationStereo · KITTI 2015 | D1: 4.32 -> 3.25 | Improved |
| MSE instead of cosine similarity distillation loss (ablation) | Fast-FoundationStereo · Middlebury-H | BP-2: 2.29 -> 2.20 | No clear effect |
| add pseudo-labeled in-the-wild data (ablation) | Fast-FoundationStereo · Middlebury-H | BP-2: 2.53 -> 2.20 | Improved |
| add pseudo-labeled data (ablation) | LightStereo-L · ETH3D | BP-1: 45.46 -> 21.12 | Improved |
| add pseudo-labeled data (ablation) | RT-IGEV · KITTI 2015 | D1: 6.00 -> 4.40 | Improved |
| use blockwise NAS instead of random assembly under same latency budget (ablation) | cost filtering module · Middlebury-Q | zero-shot error | Improved |
| prune refinement module and retrain (ablation) | refinement module · Middlebury-Q | accuracy | Improved |
| use Fast-FoundationStereo instead of FoundationStereo | FoundationStereo · Middlebury-Q | runtime (ms): 496 -> 49 | Improved |
| use Fast-FoundationStereo instead of FoundationStereo | FoundationStereo · ETH3D | BP-1: 0.50 -> 1.22 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- FoundationStereo — FoundationStereo: Zero-Shot Stereo Matching (CVPR 2025) · better on 1 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FoundationStereo: Zero-Shot Stereo Matching (CVPR 2025 · similarity 0.95)
- ZeroStereo: Zero-shot Stereo Matching from Single Images (ICCV 2025 · similarity 0.91)
- PromptStereo: Zero-Shot Stereo Matching via Structure and Motion Prompts (CVPR 2026 · similarity 0.90)
- Any-Stereo: Arbitrary Scale Disparity Estimation for Iterative Stereo Matching (AAAI 2024 · similarity 0.89)
- PPMStereo: Pick-and-Play Memory Construction for Consistent Dynamic Stereo Matching (NeurIPS 2025 · similarity 0.88)
- Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
