Research map / Detection and segmentation / Tracking
ReTracker: Exploring Image Matching for Robust Online Any Point Tracking
Dongli Tan, Xingyi He, Sida Peng, Yiqing Gong, Xing Zhu, Jiaming Sun, Ruizhen Hu, Yujun Shen, Hujun Bao, Xiaowei Zhou
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 global tracking module (ablation) | ReTracker · Ego-Retrack | AJ: 31.6 -> 44.3 | Improved |
| add two-view matching pretraining of the decoder (ablation) | ReTracker · Ego-Retrack | AJ: 41.6 -> 44.3 | Improved |
| add two-view matching pretraining of the decoder (ablation) | ReTracker · TAP-Vid DAVIS | AJ: 62.7 -> 63.5 | Improved |
| add temporal attention with history tokens (ablation) | ReTracker · Ego-Retrack | AJ: 39.3 -> 44.3 | Improved |
| add temporal attention with history tokens (ablation) | ReTracker · TAP-Vid DAVIS | AJ: 62.6 -> 63.5 | Improved |
| increase refinement history length from 3 to 12 (ablation) | ReTracker · Ego-Retrack | AJ: 40.8 -> 44.3 | Improved |
| increase history length from 12 to 15 (ablation) | ReTracker · Ego-Retrack | AJ: 44.3 -> 44.1 | No clear effect |
| increase multi-scale refinement levels from 1 to 3 (ablation) | ReTracker · Ego-Retrack | AJ: 21.1 -> 44.3 | Improved |
| increase multi-scale refinement levels from 1 to 3 (ablation) | ReTracker · TAP-Vid DAVIS | AJ: 36.3 -> 63.5 | Improved |
| restrict CoTracker3 to online mode (no future frames, step size 1) | CoTracker3 · Ego-Retrack | AJ: 46.0 -> 27.6 | Hurt |
Closest papers by meaning
- Track-On2: Enhancing Online Point Tracking With Memory (TPAMI 2026 · similarity 0.89)
- GlobalTrack: A Simple and Strong Baseline for Long-Term Tracking (AAAI 2020 · similarity 0.89)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.88)
- Effective Local and Global Search for Fast Long-Term Tracking (TPAMI 2022 · similarity 0.88)
- SMILEtrack: SiMIlarity LEarning for Occlusion-Aware Multiple Object Tracking (AAAI 2024 · similarity 0.88)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 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.
