Research map / Image generation / Image restoration
BlinkTrack: Feature Tracking over 80 FPS via Events and Images
Yichen Shen, Yijin Li, Shuo Chen, Guanglin Li, Zhaoyang Huang, Hujun Bao, Zhaopeng Cui, Guofeng Zhang
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 the feature pyramid encoder (ablation) | BlinkTrack event module · EC | expected feature age: 0.789 -> 0.819 | Improved |
| add the feature pyramid encoder (ablation) | BlinkTrack event module · EC-occ | expected feature age: 0.439 -> 0.522 | Improved |
| add the feature pyramid encoder (ablation) | BlinkTrack event module · EDS | expected feature age: 0.395 -> 0.474 | Improved |
| add both LSTM modules (feature and displacement) (ablation) | BlinkTrack event module · EC | expected feature age: 0.568 -> 0.819 | Improved |
| add the displacement LSTM to the feature LSTM (ablation) | BlinkTrack event module · EC-occ | expected feature age: 0.473 -> 0.522 | Improved |
| add the feature LSTM to the displacement LSTM (ablation) | BlinkTrack event module · EDS | expected feature age: 0.282 -> 0.474 | Improved |
| take correlation vector from feature map center instead of U-Net bottleneck (ablation) | BlinkTrack event module · EC-occ | expected feature age: 0.472 -> 0.522 | Improved |
| add the learnable Kalman filter to event+image fusion | BlinkTrack E+I · EC-occ | delta_avg occluded: 11.4 -> 28.5 | Improved |
| add the learnable Kalman filter to event+image fusion | BlinkTrack E+I · EDS-occ | delta_avg all: 41.6 -> 50.6 | Improved |
| add the Kalman filter to event-only module | BlinkTrack event module · EC | expected feature age: 0.819 -> 0.820 | No clear effect |
| add the Kalman filter to image-only module | BlinkTrack image module · EDS | expected feature age: 0.524 -> 0.511 | Hurt |
| naively fuse event tracker with KLT via initial position replacement | Deep-EV-Tracker · EC | expected feature age: 0.787 -> 0.730 | Hurt |
Closest papers by meaning
- Data-Driven Feature Tracking for Event Cameras With and Without Frames (TPAMI 2025 · similarity 0.90)
- Data-Driven Feature Tracking for Event Cameras (CVPR 2023 · similarity 0.89)
- "BlinkVision: A Benchmark for Optical Flow, Scene Flow and Point Tracking Estimation using RGB Frames and Events" (ECCV 2024 · similarity 0.89)
- ETAP: Event-based Tracking of Any Point (CVPR 2025 · similarity 0.89)
- TimeTracker: Event-based Continuous Point Tracking for Video Frame Interpolation with Non-linear Motion (CVPR 2025 · similarity 0.87)
- CSTrack: Enhancing RGB-X Tracking via Compact Spatiotemporal Features (ICML 2025 · 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.
