Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Learning the Distribution of Errors in Stereo Matching for Joint Disparity and Uncertainty Estimation
Read the paperCheck 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 |
|---|---|---|---|
| Add KL divergence matching the predicted uncertainty and disparity-error distributions | GwcNet with the log-likelihood uncertainty loss · Scene Flow | EPE: 0.7611 -> 0.6754 | Helped |
| Add KL divergence matching the predicted uncertainty and disparity-error distributions | GwcNet with the log-likelihood uncertainty loss · VK2-S6 | AUC Est.: 12.5320 -> 9.7637 | Helped |
| Add KL divergence matching the predicted uncertainty and disparity-error distributions | GwcNet with the log-likelihood uncertainty loss · DrivingStereo test (FT) | EPE: 0.5332 -> 0.5264 | Helped |
| Add KL divergence matching the predicted uncertainty and disparity-error distributions | GwcNet with the log-likelihood uncertainty loss · DS-Weather | EPE: 2.3944 -> 1.5637 | Helped |
| Use logarithmically spaced rather than linearly spaced histogram bins (ablation) | SEDNet · Scene Flow | EPE: 0.6827 -> 0.6754 | Helped |
| Increase histogram bins from 11 to 20 (ablation) | SEDNet · Scene Flow | EPE: 0.6754 -> 0.6762 | No clear effect |
| Increase histogram bins from 11 to 50 (ablation) | SEDNet · Scene Flow | EPE: 0.6754 -> 0.6894 | Hurt |
| Use an adaptive inlier threshold μ_ε+3b_ε instead of fixed EPE<5 (ablation) | SEDNet · Scene Flow | EPE: 1.0046 -> 0.6754 | Helped |
| Set the adaptive inlier threshold to μ_ε+3b_ε rather than μ_ε+1b_ε (ablation) | SEDNet · VK2-S6 | EPE: 0.3896 -> 0.3109 | Helped |
| Set the adaptive inlier threshold to μ_ε+3b_ε rather than μ_ε+5b_ε (ablation) | SEDNet · VK2-S6 | EPE: 0.3236 -> 0.3109 | Helped |
| Set the adaptive inlier threshold to μ_ε+3b_ε rather than μ_ε+1b_ε (ablation) | SEDNet · DS-Weather | EPE: 6.9335 -> 1.5637 | Helped |
| Set the adaptive inlier threshold to μ_ε+5b_ε rather than μ_ε+3b_ε (ablation) | SEDNet · DS-Weather | EPE: 1.5637 -> 1.7051 | Hurt |
Closest papers by meaning
- Stereo Risk: A Continuous Modeling Approach to Stereo Matching (ICML 2024 · similarity 0.91)
- Modeling Stereo-Confidence out of the End-to-End Stereo-Matching Network via Disparity Plane Sweep (AAAI 2024 · similarity 0.90)
- CFNet: Cascade and Fused Cost Volume for Robust Stereo Matching (CVPR 2021 · similarity 0.89)
- Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail (CVPR 2025 · similarity 0.87)
- Adaptive Multi-Modal Cross-Entropy Loss for Stereo Matching (CVPR 2024 · similarity 0.87)
- Robust Synthetic-to-Real Transfer for Stereo Matching (CVPR 2024 · 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.