Research map / Label-efficient and robust learning / Recognition and long-tail
Optical Flow Requires Multiple Strategies (but Only One Network)
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 |
|---|---|---|---|
| Replace PatchBatch with the improved PatchBatch+SPCI system | Optical flow on KITTI2012 · KITTI2012 | Out-Noc: 5.44% -> 4.65% | Helped |
| Use improved PatchBatch+SPCI instead of PatchBatch | Optical flow on KITTI2015 · KITTI2015 | Fl-all: 21.69% -> 18.46% | Helped |
| Use SPCI rather than interleaving alone (ablation) | Improved PatchBatch on MPI-Sintel · MPI-Sintel | EPE: 6.22 -> 6.24 | Hurt |
| Add interleaved negative sampling (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.85% -> 4.41% | Helped |
| Add the SPCI curriculum and self-paced difficulty controls to interleaving (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.41% -> 4.37% | Helped |
| Use interleaving instead of self-paced learning alone (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 5.23% -> 4.41% | Helped |
| Invert the displacement-to-negative-distance relationship (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.85% -> 8.30% | Hurt |
| Add the standard-deviation term to Hinge loss (ablation) | Hinge PatchBatch · KITTI2012 validation | error percent post EF: 5.18% -> 4.85% | Helped |
| Replace CENT loss with Hinge loss (ablation) | PatchBatch descriptor training · KITTI2012 validation | error percent post PM: 9.93% -> 7.78% | Helped |
| Increase patch size to 71×71 pixels (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.37% -> 4.27% | Helped |
| Use negative mining (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.85% -> 5.00% | Hurt |
| Use curriculum by displacement instead of standard Hinge+SD training (ablation) | Hinge+SD PatchBatch · KITTI2012 validation | error percent post EF: 4.85% -> 4.83% | No clear effect |
Closest papers by meaning
- PatchBatch: A Batch Augmented Loss for Optical Flow (CVPR 2016 · similarity 0.88)
- CNN-Based Patch Matching for Optical Flow With Thresholded Hinge Embedding Loss (CVPR 2017 · similarity 0.88)
- Rethinking Optical Flow From Geometric Matching Consistent Perspective (CVPR 2023 · similarity 0.87)
- A Theory on Flow Matching with Neural Networks (NeurIPS 2026 · similarity 0.86)
- Context-Aware Iteration Policy Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.86)
- Contextual Flow Matching: Adaptive Step Selection in Flow Models for Efficient Visual Generation (NeurIPS 2026 · similarity 0.85)
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.