Research map / 3D vision / Depth and stereo
A Continuous Optimization Approach for Efficient and Accurate Scene Flow
Zhaoyang Lv, Chris Beall, Pablo F. Alcantarilla, Fuxin Li, Zsolt Kira, Frank Dellaert
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 |
|---|---|---|---|
| add matching factor to census factor (ablation) | stereo estimation · KITTI training stereo (NOC) | D1-all error: 12.31 -> 6.93 | Improved |
| add continuity factor to census factor (ablation) | stereo estimation · KITTI training stereo (NOC) | D1-all error: 12.31 -> 5.94 | Improved |
| use all stereo factors instead of census + matching (ablation) | stereo estimation · KITTI training stereo (NOC) | D1-all error: 6.93 -> 4.85 | Improved |
| add census factor to matching factor (ablation) | stereo estimation · KITTI training stereo (NOC) | D1-all error: 7.62 -> 6.93 | Improved |
| use multi-scale census instead of raw census (ablation) | flow estimation · KITTI training flow (NOC) | F-all error: 14.20 -> 12.45 | Improved |
| use census + continuity instead of raw census only (ablation) | flow estimation · KITTI training flow (NOC) | F-all error: 14.20 -> 12.44 | Improved |
| add all factors instead of multi-scale census alone (ablation) | flow estimation · KITTI training flow (NOC) | F-all error: 12.45 -> 12.31 | Improved |
| use matching only instead of raw census (ablation) | flow estimation · KITTI training flow (NOC) | F-all error: 14.20 -> 13.20 | Improved |
| use proposed method instead of OSF | scene flow · KITTI Scene Flow test (OCC) | runtime: 50 min -> 80 s | Improved |
| use proposed method instead of PRSM | scene flow · KITTI Scene Flow test (OCC) | SF error all: 9.44 -> 16.33 | Hurt |
Closest papers by meaning
- Fast Multi-Frame Stereo Scene Flow With Motion Segmentation (CVPR 2017 · similarity 0.90)
- Optical Flow in Mostly Rigid Scenes (CVPR 2017 · similarity 0.88)
- SceneFlowFields++: Multi-frame Matching, Visibility Prediction, and Robust Interpolation for Scene Flow Estimation (IJCV 2019 · similarity 0.88)
- SEA-Flow3D: Simplified, Efficient, and Accurate Scene Flow via Spatial Vector Sampling and Multi-scale Refinement (CVPR 2026 · similarity 0.88)
- Consistency Guided Scene Flow Estimation (ECCV 2020 · similarity 0.88)
- DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Iterative Diffusion-Based Refinement (CVPR 2024 · similarity 0.86)
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.
