Research map / Architectures and efficiency / RNNs and normalization
Guided Stereo Matching
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Apply sparse guidance at test time to pretrained iResNet | iResNet · KITTI 2015 | average error (px): 1.863 -> 1.431 | Helped |
| Train iResNet with sparse guidance (ablation) | iResNet · KITTI 2015 | average error (px): 1.863 -> 0.904 | Helped |
| Fine-tune and train iResNet with sparse guidance (ablation) | iResNet-ft · KITTI 2015 | average error (px): 1.049 -> 0.735 | Helped |
| Apply sparse guidance at test time to pretrained PSMNet | PSMNet · KITTI 2015 | average error (px): 4.716 -> 3.900 | Helped |
| Train PSMNet with sparse guidance (ablation) | PSMNet · KITTI 2015 | average error (px): 4.716 -> 1.395 | Helped |
| Apply sparse guidance at test time | iResNet and PSMNet · ETH3D | average error (px): 5.120 -> 2.717; 5.078 -> 4.739 | Helped |
| Train the network with sparse guidance (ablation) | PSMNet · Middlebury v3 additional | error rate >0.5: 54.717 -> 83.433 | Hurt |
| Train the network with sparse guidance (ablation) | PSMNet · Middlebury v3 additional | error rate >2: 20.239 -> 7.472 | Helped |
| Apply sparse guidance at test time to SGM | SGM · KITTI 2012 | average error (px): 2.740 -> 2.153 | Helped |
| Use Gaussian modulation with k=10 and c=1 instead of no modulation (ablation) | iResNet · KITTI 2015 | average error (px): 1.863 -> 1.338 | Helped |
| Increase Gaussian modulation magnitude to k=100 at c=1 (ablation) | iResNet · KITTI 2015 | average error (px): 1.863 -> 6.857 | Hurt |
| Use Gaussian modulation with k=10 and c=1 instead of no modulation (ablation) | PSMNet · KITTI 2015 | average error (px): 4.716 -> 3.900 | Helped |
Closest papers by meaning
- Zoom and Learn: Generalizing Deep Stereo Matching to Novel Domains (CVPR 2018 · similarity 0.91)
- Fast-MVSNet: Sparse-to-Dense Multi-View Stereo With Learned Propagation and Gauss-Newton Refinement (CVPR 2020 · similarity 0.90)
- CFNet: Cascade and Fused Cost Volume for Robust Stereo Matching (CVPR 2021 · similarity 0.90)
- Stereo Risk: A Continuous Modeling Approach to Stereo Matching (ICML 2024 · similarity 0.89)
- Improved Stereo Matching With Constant Highway Networks and Reflective Confidence Learning (CVPR 2017 · similarity 0.89)
- AANet: Adaptive Aggregation Network for Efficient Stereo Matching (CVPR 2020 · similarity 0.89)
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.