Research map / 3D vision / Depth and stereo
"BlinkVision: A Benchmark for Optical Flow, Scene Flow and Point Tracking Estimation using RGB Frames and Events"
Yijin Li, Yichen Shen, Zhaoyang Huang, Shuo Chen, Weikang Bian, Xiaoyu Shi, Fu-Yun Wang, Keqiang Sun, Hujun Bao, Zhaopeng Cui, Guofeng Zhang, Hongsheng Li
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 |
|---|---|---|---|
| fine-tune on BlinkVision training set | FlowFormer++ · BlinkVision RGB clean optical flow | EPE avg: 7.20 -> 4.73 | Improved |
| increase frame stride from 1 to 8 | RAFT · BlinkVision RGB clean optical flow | EPE: 1.83 -> 20.46 | Hurt |
| use final RGB (blur, limited dynamic range) instead of clean RGB | RGB optical flow methods · BlinkVision optical flow | EPE: 7.20 -> 8.90 (FF++ avg) | Hurt |
| use event-only methods instead of RGB methods without fine-tuning | E-RAFT vs RAFT · BlinkVision optical flow | EPE avg: 9.27 (RAFT) vs 14.07 (E-RAFT) | Hurt |
| fine-tune event methods on BlinkVision events | E-FlowFormer · BlinkVision event optical flow | EPE avg: 12.76 -> 6.02 | Improved |
| track ORB feature points instead of grid-sampled points (ablation) | HASTE · BlinkVision event point tracking | delta_avg: 18.60 -> 27.65 | Improved |
| increase stride from 1 to 2 | PIPs · BlinkVision RGB clean point tracking | delta_avg: 37.77 -> 39.34 | Improved |
| fine-tune on BlinkVision then test on Sintel clean | RAFT · Sintel (clean) | EPE: 2.08 -> 1.69 | Improved |
| fine-tune on BlinkVision then test on FlyingObjects | DCEIFlow · FlyingObjects | EPE: 9.58 -> 2.37 | Improved |
| fine-tune on BlinkVision then test on KITTI | RAFT · KITTI | EPE: 5.10 -> 4.66 | Improved |
Closest papers by meaning
- BlinkTrack: Feature Tracking over 80 FPS via Events and Images (ICCV 2025 · similarity 0.89)
- AllTracker: Efficient Dense Point Tracking at High Resolution (ICCV 2025 · similarity 0.88)
- EMatch: A Unified Framework for Event-based Optical Flow and Stereo Matching (ICCV 2025 · similarity 0.88)
- ARES: Unifying Asymmetric RGB-Event Stereo for Probabilistic Scene Flow Estimation (CVPR 2026 · similarity 0.88)
- PointOdyssey: A Large-Scale Synthetic Dataset for Long-Term Point Tracking (ICCV 2023 · similarity 0.88)
- Dense Optical Tracking: Connecting the Dots (CVPR 2024 · similarity 0.87)
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.
