Research map / Learning theory and algorithms / Combinatorial and search optimization
The Fast Bilateral Solver
Jonathan T. Barron, Ben Poole
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 |
|---|---|---|---|
| post-process MC-CNN depth with the robust bilateral solver | MC-CNN · Middlebury v3 test | MAE (All): 17.9 -> 8.19 | Improved |
| post-process MC-CNN depth with the robust bilateral solver | MC-CNN · Middlebury v3 test | bad 1% (All): 28.1 -> 28.2 | No clear effect |
| post-process MC-CNN depth with the robust bilateral solver | MC-CNN · Middlebury v3 test | RMSE (All): 55.0 -> 29.9 | Improved |
| post-process MC-CNN depth with the robust bilateral solver | MC-CNN · Middlebury v3 train | MAE (All): 5.93 -> 2.81 | Improved |
| post-process MC-CNN depth with tree filtering | MC-CNN · Middlebury v3 train | bad 1% (All): 20.07 -> 29.15 | Hurt |
| post-process MC-CNN depth with weighted median filter | MC-CNN · Middlebury v3 train | bad 1% (All): 20.07 -> 33.37 | Hurt |
| post-process MC-CNN depth with RBS instead of domain transform | MC-CNN + DT · Middlebury v3 train | MAE (All): 5.69 -> 2.81 | Improved |
| post-process MeshStereo depth with RBS | MeshStereo · Middlebury v3 train | MAE (All): 3.83 -> 3.22 | Improved |
| post-process SGM depth with RBS | SGM · Middlebury v3 train | MAE (All): 3.85 -> 3.44 | Improved |
| post-process TMAP depth with RBS | TMAP · Middlebury v3 train | MAE (All): 3.98 -> 3.31 | Improved |
| reduce multi-channel right-hand sides via low-rank QR approximation (epsilon 0.01) (ablation) | bilateral solver · semantic segmentation | runtime: approx 5x faster | Improved |
| replace Barron et al. defocus solver with bilateral solver on reparametrized interval input | defocus stereo solver · stereo defocus | speed: 3.5x faster | Improved |
Closest papers by meaning
- The Domain Transform Solver (CVPR 2019 · similarity 0.87)
- Edge-aware Image Smoothing with Relative Wavelet Domain Representation (ICLR 2025 · similarity 0.86)
- EdgeStereo: An Effective Multi-task Learning Network for Stereo Matching and Edge Detection (IJCV 2020 · similarity 0.84)
- PatchMatch Filter: Edge-Aware Filtering Meets Randomized Search for Visual Correspondence (TPAMI 2016 · similarity 0.84)
- Stereo Risk: A Continuous Modeling Approach to Stereo Matching (ICML 2024 · similarity 0.83)
- FoundationStereo: Zero-Shot Stereo Matching (CVPR 2025 · similarity 0.83)
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.
