Research map / Detection and segmentation / Tracking
DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 OC-SORT association instead of IoU matching | YOLOX detections · DanceTrack val | HOTA: 44.7 -> 52.1 | Helped |
| Use LSTM motion instead of no temporal motion model | YOLOX detections with IoU association · DanceTrack val | HOTA: 44.7 -> 51.6 | Helped |
| Add COCO mask training data (ablation) | CenterNet + BYTE · DanceTrack val | HOTA: 36.9 -> 38.1 | Helped |
| Add mask output for association (ablation) | CenterNet + BYTE trained with COCO masks · DanceTrack val | HOTA: 38.1 -> 39.2 | Helped |
| Add COCO pose training data (ablation) | CenterNet + BYTE · DanceTrack val | HOTA: 36.9 -> 40.6 | Helped |
| Add pose output for association (ablation) | CenterNet + BYTE trained with COCO pose · DanceTrack val | HOTA: 40.6 -> 41.0 | Helped |
| Add KITTI training data (ablation) | CenterNet + BYTE · DanceTrack val | HOTA: 36.9 -> 34.4 | Hurt |
| Add depth output for association (ablation) | CenterNet + BYTE trained jointly on DanceTrack and KITTI · DanceTrack val | HOTA: 34.4 -> 35.1 | Helped |
| Add Kalman filter motion to IoU matching (ablation) | Ground-truth boxes · DanceTrack oracle val | HOTA: 72.8 -> 69.4 | Hurt |
| Add appearance matching to IoU and motion (ablation) | Ground-truth boxes · DanceTrack oracle val | HOTA: 69.4 -> 59.7 | Hurt |
| Use appearance matching without IoU or motion (ablation) | Ground-truth boxes · DanceTrack oracle val | HOTA: 72.8 -> 68.0 | Hurt |
| Add appearance matching to SORT | SORT with YOLOX detections · DanceTrack val | HOTA: 47.8 -> 45.8 | Hurt |
Closest papers by meaning
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.90)
- Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024 · similarity 0.89)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.89)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.89)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.89)
- UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation (AAAI 2024 · similarity 0.88)
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.