Research map / 3D vision / Depth and stereo
Learning Optical Expansion From Scale Matching
Han Ling, Yinghui Sun, Quansen Sun, Zhenwen Ren
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 |
|---|---|---|---|
| Add TPCV scale correlation to RAFT | RAFT · KITTI test | Fl-fg outlier rate: 6.87 -> 5.52 | Improved |
| Add TPCV scale correlation to RAFT | RAFT · KITTI test | Fl-all outlier rate: 5.1 -> 4.69 | Improved |
| Estimate motion-in-depth through TPCV scale matching | RAFT · K-40 | MIDerror: 73.55 -> 42.84 | Improved |
| Estimate TTC from TPCV motion-in-depth | RAFT · K-40 | Err-1s: 1.86 -> 1.14 | Improved |
| Add TPCV to RAFT3D correlation and estimate depth-change ratio | RAFT3D · KITTI scene flow benchmark | SF-all: 5.77 -> 5.00 | Improved |
| Add TPCV to RAFT3D correlation and estimate depth-change ratio | RAFT3D · KITTI scene flow benchmark | Fl-fg: 8.79 -> 10.19 | Hurt |
| Replace TPCV with direct scale-flow regression (ablation) | RAFT+TPCV · K-40 | Flowepe: 1.31 -> 1.61 | Hurt |
| Replace TPCV with direct scale-flow regression (ablation) | RAFT+TPCV · K-40 | MIDerror: 42.8 -> 56.2 | Hurt |
| Use direct image-pyramid matching instead of TPCV (ablation) | RAFT+TPCV · K-40 | Flowepe: 1.31 -> 1.42 | Hurt |
| Use direct image-pyramid matching instead of TPCV (ablation) | RAFT+TPCV · K-40 | MIDerror: 42.8 -> 48.9 | Hurt |
| Apply TPCV to RAFT3D | RAFT3D · FlyingThings3D | end-to-end 3D error: 0.062m to 0.048m | Improved |
| Apply TPCV to RAFT3D | RAFT3D · FlyingThings3D | end-to-end 2D error: 2.2px to 1.95px | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 3 on most shared results.
- CamLiFlow — CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint Optical Flow and Scene Flow Estimation (CVPR 2022) · better on 0 of 1 shared result
- Binary TTC — Binary TTC: A Temporal Geofence for Autonomous Navigation (CVPR 2021) · better on 1 of 1 shared result
- RAFT3D — RAFT-3D: Scene Flow Using Rigid-Motion Embeddings (CVPR 2021) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Digging Into Monocular Texture-Scale Feature Fusion for Optical Flow and Optical Expansion Estimation (TPAMI 2026 · similarity 0.88)
- Upgrading Optical Flow to 3D Scene Flow Through Optical Expansion (CVPR 2020 · similarity 0.88)
- Optical Flow Matching: Reframing Optical Flow as Continuous Transport Dynamics (CVPR 2026 · similarity 0.84)
- Optical Flow as Spatial-Temporal Attention Learners (TPAMI 2024 · similarity 0.83)
- Structure-from-Motion-Aware PatchMatch for Adaptive Optical Flow Estimation (ECCV 2018 · similarity 0.83)
- DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (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.
