Research map / Detection and segmentation / Tracking
Tracking by Associating Clips
Sanghyun Woo, Kwanyong Park, Seoung Wug Oh, In So Kweon, Joon-Young Lee
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 |
|---|---|---|---|
| replace frame-wise QDTrack matching with Clip Tracker clip-wise association | QDTrack · TAO val | Track AP50: 13.4 -> 17.7 | Improved |
| replace frame-wise QDTrack matching with Clip Tracker clip-wise association | QDTrack · TAO test | Track AP50: 12.6 -> 15.8 | Improved |
| replace frame-wise QDTrack matching with Clip Tracker clip-wise association | QDTrack · MOT17 val (half split) | IDF1: 69.5 -> 72.4 | Improved |
| replace frame-wise QDTrack matching with Clip Tracker clip-wise association | QDTrack · MOT17 test | HOTA: 57.8 -> 59.0 | Improved |
| simple clip-based tracker with temporal averaged features instead of frame-based (ablation) | QDTrack · TAO val | Track AP50: 13.4 -> 12.3 | Hurt |
| simple clip-based tracker with temporal averaged features instead of frame-based (ablation) | QDTrack · MOT17 val (half split) | HOTA: 57.5 -> 58.6 | Improved |
| use Clip Tracker instead of temporal averaged feature matching for inter-clip (ablation) | Tmp avg. · TAO val | Track AP50: 12.3 -> 17.7 | Improved |
| use Clip Tracker instead of temporal averaged feature matching for inter-clip (ablation) | Tmp avg. · MOT17 val (half split) | HOTA: 58.6 -> 59.7 | Improved |
| use IoU-based chaining instead of feature matching for inter-clip (ablation) | Tmp avg. · TAO val | Track AP50: 12.3 -> 10.2 | Hurt |
| use direction-free hierarchical clustering instead of directional matching for intra-clip (ablation) | Directional matching · MOT17 val (half split) | HOTA: 58.6 -> 57.9 | Mixed |
| add overlap between clips (interval 5 instead of 10 at size 10) (ablation) | Clip Tracker · MOT17 val (half split) | HOTA: 58.6 -> 59.7 | Improved |
| increase clip size from 10 to 20 (interval 10 vs 5) (ablation) | Clip Tracker · MOT17 val (half split) | HOTA: 59.7 -> 59.5 | Hurt |
Closest papers by meaning
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.88)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.87)
- TrackIME: Enhanced Video Point Tracking via Instance Motion Estimation (NeurIPS 2024 · similarity 0.87)
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.87)
- ADA-Track: End-to-End Multi-Camera 3D Multi-Object Tracking with Alternating Detection and Association (CVPR 2024 · similarity 0.87)
- A Distractor-Aware Memory for Visual Object Tracking with SAM2 (CVPR 2025 · 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.
