Research map / Video / Video segmentation and tracking
AccFlow: Backward Accumulation for Long-Range Optical Flow
Guangyang Wu, Xiaohong Liu, Kunming Luo, Xi Liu, Qingqing Zheng, Shuaicheng Liu, Xinyang Jiang, Guangtao Zhai, Wenyi Wang
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 |
|---|---|---|---|
| use backward instead of forward accumulation in AccPlus (with adaptive blending) (ablation) | Acc+RAFT · HS-Sintel | EPE ALL: 1.716 -> 1.383 | Improved |
| use backward instead of forward accumulation (without AB) (ablation) | Acc+RAFT · CVO (Final) | EPE ALL: 3.328 -> 2.709 | Improved |
| add adaptive blending module to backward accumulation (ablation) | Acc+RAFT · HS-Sintel | EPE ALL: 1.740 -> 1.383 | Improved |
| add adaptive blending module to backward accumulation (ablation) | Acc+RAFT · CVO (Final) | EPE ALL: 2.709 -> 2.707 | No clear effect |
| accumulate flows via AccFlow instead of direct estimation with fine-tuned RAFT | RAFT* · CVO (Final) | EPE ALL: 4.537 -> 2.707 | Improved |
| use explicit forward accumulation (Lim) instead of direct estimation | RAFT* · CVO (Final) | EPE ALL: 4.537 -> 13.02 | Hurt |
| use warm start (implicit forward accumulation) | RAFT* · CVO (Final) | EPE ALL: 4.537 -> 3.808 | Improved |
| fine-tune the optical flow estimator inside AccFlow instead of freezing it | Acc+RAFT · CVO (Clean) | EPE ALL: 3.170 -> 2.634 | Improved |
| AccFlow versus direct RAFT inference cost | Acc+RAFT* · CVO inference time | inference time (s): 0.129 -> 0.813 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SlowFlow — Slow Flow: Exploiting High-Speed Cameras for Accurate and Diverse Optical Flow Reference Data (CVPR 2017) · better on 1 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
- Learning Large Motion Estimation from Intermediate Representations with a High-Resolution Optical Flow Dataset Featuring Long-Range Dynamic Motion (ICCV 2025 · similarity 0.88)
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.86)
- DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (CVPR 2025 · similarity 0.86)
- MemFlow: Optical Flow Estimation and Prediction with Memory (CVPR 2024 · similarity 0.86)
- StreamFlow: Streamlined Multi-Frame Optical Flow Estimation for Video Sequences (NeurIPS 2024 · similarity 0.86)
- OCAI: Improving Optical Flow Estimation by Occlusion and Consistency Aware Interpolation (CVPR 2024 · 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.
