Research map / Detection and segmentation / 3D detection and driving
Tracking Objects as Points
Xingyi Zhou, Vladlen Koltun, Philipp Krähenbühl
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 prior-frame heatmap input (ablation) | CenterTrack · MOT17 val | MOTA: 63.9 -> 66.1 | Improved |
| add prior-frame heatmap input (ablation) | CenterTrack · KITTI val | MOTA: 85.4 -> 88.7 | Improved |
| add prior-frame heatmap input (ablation) | CenterTrack · nuScenes val | AMOTA@0.2: 26.5 -> 28.3 | Improved |
| add learned offset prediction (ablation) | CenterTrack · MOT17 val | MOTA: 65.8 -> 66.1 | Improved |
| add learned offset prediction (ablation) | CenterTrack · KITTI val | MOTA: 87.1 -> 88.7 | Improved |
| add learned offset prediction (ablation) | CenterTrack · nuScenes val | AMOTA@0.2: 17.8 -> 28.3 | Improved |
| add heatmap noise simulation during training (ablation) | CenterTrack · MOT17 val | MOTA: 34.4 -> 66.1 | Improved |
| add heatmap noise simulation during training (ablation) | CenterTrack · KITTI val | MOTA: 80.1 -> 88.7 | Improved |
| train on static images instead of video (ablation) | CenterTrack · MOT17 val | MOTA: 66.1 -> 66.1 | No clear effect |
| train on static images instead of video (ablation) | CenterTrack · KITTI val | MOTA: 88.7 -> 86.8 | Hurt |
| use Hungarian matching instead of greedy matching (ablation) | CenterTrack · MOT17 val | MOTA: 66.1 -> 66.1 | No clear effect |
| add track rebirth (K=32) (ablation) | CenterTrack · MOT17 val | IDF1: 64.2 -> 69.4 | Improved |
Closest papers by meaning
- Center-Based 3D Object Detection and Tracking (CVPR 2021 · similarity 0.90)
- OmniTracker: Unifying Visual Object Tracking by Tracking-With-Detection (TPAMI 2025 · similarity 0.89)
- VastTrack: Vast Category Visual Object Tracking (NeurIPS 2024 · similarity 0.88)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.88)
- UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation (AAAI 2024 · similarity 0.88)
- TrackIME: Enhanced Video Point Tracking via Instance Motion Estimation (NeurIPS 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.
