Research map / 3D vision / Depth and stereo
Structure-from-Motion-Aware PatchMatch for Adaptive Optical Flow Estimation
Daniel Maurer, Nico Marniok, Bastian Goldluecke, Andres Bruhn
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 |
|---|---|---|---|
| replace CPM matches with three-frame structure matching (temporal averaging) (ablation) | baseline · KITTI 2012 train | AEE: 2.61 -> 1.25 | Improved |
| replace CPM matches with three-frame structure matching (temporal averaging) (ablation) | baseline · KITTI 2015 train | AEE: 6.82 -> 7.85 | Hurt |
| replace CPM matches with three-frame structure matching (temporal averaging) (ablation) | baseline · MPI Sintel train (clean) | AEE: 1.95 -> 20.68 | Hurt |
| use three frames instead of two for structure matching (temporal selection) (ablation) | structure matching · MPI Sintel train (clean) | AEE: 17.09 -> 15.69 | Improved |
| use three frames instead of two for structure matching (temporal averaging) (ablation) | structure matching · KITTI 2012 train | AEE: 2.25 -> 1.25 | Improved |
| add backward matches to forward matches (ablation) | baseline · KITTI 2012 train | AEE: 2.61 -> 4.53 | Hurt |
| combine structure matches permissively instead of baseline only (ablation) | baseline · MPI Sintel train (clean) | AEE: 1.95 -> 2.92 | Hurt |
| combine structure matches restrictively with consistency check (ablation) | baseline · KITTI 2012 train | AEE: 2.61 -> 1.59 | Improved |
| add voting rule for non-checkable structure matches over restrictive (temporal averaging) (ablation) | restrictive approach · KITTI 2015 train | AEE: 5.04 -> 4.16 | Improved |
| add voting rule for non-checkable structure matches over restrictive (temporal averaging) (ablation) | restrictive approach · MPI Sintel train (clean) | AEE: 1.90 -> 1.92 | No clear effect |
| voting approach versus baseline (ablation) | CPM baseline · KITTI 2015 train | AEE: 6.82 -> 4.16 | Improved |
| use a single refinement parameter set instead of per-benchmark optimization (ablation) | voting approach (temporal averaging) · KITTI 2012 train | AEE: 1.30 -> 1.31 | No clear effect |
Closest papers by meaning
- Explicit Motion Disentangling for Efficient Optical Flow Estimation (ICCV 2023 · similarity 0.89)
- Efficient Coarse-To-Fine PatchMatch for Large Displacement Optical Flow (CVPR 2016 · similarity 0.89)
- Optical Flow Matching: Reframing Optical Flow as Continuous Transport Dynamics (CVPR 2026 · similarity 0.89)
- DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (CVPR 2025 · similarity 0.87)
- Injecting Frame-Event Complementary Fusion into Diffusion for Optical Flow in Challenging Scenes (NeurIPS 2025 · similarity 0.87)
- 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.
