Research map / 3D vision / Depth and stereo
Using Self-Contradiction to Learn Confidence Measures in Stereo Vision
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 |
|---|---|---|---|
| Use automatically generated labels instead of laser labels for training | Ensemble, Park, and GCP confidence measures with SGM · KITTI2012 | AUSC: 1.97 -> 1.95; 1.93 -> 1.92; 2.50 -> 2.45 | Helped |
| Use automatically generated labels instead of laser labels for training | Ensemble, Park, and GCP confidence measures with SPS · KITTI2012 | AUSC: 5.86 -> 5.43; 6.23 -> 5.61; 8.28 -> 7.95 | Helped |
| Train on automatically generated labels from the target dataset rather than KITTI labels | Ensemble, Park, and GCP confidence measures with SGM · Middlebury2014 | AUSC: 1.24 -> 1.19; 1.25 -> 1.20; 1.78 -> 1.50 | Helped |
| Train on automatically generated labels from the target dataset rather than KITTI labels | Ensemble, Park, and GCP confidence measures with SPS · Middlebury2014 | AUSC: 1.48 -> 1.39; 1.81 -> 1.42; 2.05 -> 1.44 | Helped |
| Add automatically generated scene-specific training labels | Ensemble, Park, and GCP confidence measures · Strecha Fountain | Relative area (RA): 1.43 -> 1.40; 1.59 -> 1.51; 2.60 -> 2.01 | Helped |
| Use Middlebury-trained measures instead of KITTI-trained measures | Ensemble, Park, and GCP confidence measures · Strecha Fountain | Relative area (RA): 1.81 -> 1.43; 1.91 -> 1.59; 3.54 -> 2.60 | Helped |
| Use automatically generated labels from only 7 available sequences rather than labels from 195 testing sequences (ablation) | All three confidence prediction approaches · KITTI2012 | AUSC: increased by ~10% | Hurt |
| Increase automatically generated training data volume and diversity rather than relying on a limited laser-labeled set | All three confidence prediction approaches · KITTI2012 | AUSC: 1 to 10% | Helped |
Closest papers by meaning
- Learning to Predict Stereo Reliability Enforcing Local Consistency of Confidence Maps (CVPR 2017 · similarity 0.93)
- Modeling Stereo-Confidence out of the End-to-End Stereo-Matching Network via Disparity Plane Sweep (AAAI 2024 · similarity 0.89)
- LAF-Net: Locally Adaptive Fusion Networks for Stereo Confidence Estimation (CVPR 2019 · similarity 0.86)
- On the Uncertainty of Self-Supervised Monocular Depth Estimation (CVPR 2020 · similarity 0.86)
- Learning the Distribution of Errors in Stereo Matching for Joint Disparity and Uncertainty Estimation (CVPR 2023 · similarity 0.85)
- Stereo Risk: A Continuous Modeling Approach to Stereo Matching (ICML 2024 · similarity 0.84)
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.