Research map / Point clouds / Segmentation and detection
Segment as Points for Efficient Online Multi-Object Tracking and Segmentation
Zhenbo Xu, Wei Zhang, Xiao Tan, Wei Yang, Huan Huang, Shilei Wen, Errui Ding, Liusheng Huang
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 temporal consistency loss (ablation) | PointTrack · KITTI MOTS val (cars) | sMOTSA: 82.9 -> 85.5 | Improved |
| add temporal consistency loss (ablation) | PointTrack · KITTI MOTS val (pedestrians) | sMOTSA: 61.4 -> 62.4 | Improved |
| extract features from segments instead of bbox-sampled points (ablation) | PointTrack · KITTI MOTS val (cars) | IDS: 36 -> 22 | Improved |
| extract features from segments instead of bbox-sampled points (ablation) | PointTrack · KITTI MOTS val (pedestrians) | sMOTSA: 61.8 -> 62.4 | Improved |
| use all four data modalities instead of color only (ablation) | PointTrack · KITTI MOTS val (cars) | IDS: 171 -> 22 | Improved |
| add color data to offset, category, position (ablation) | PointTrack · KITTI MOTS val (cars) | sMOTSA: 83.65 -> 85.51 | Improved |
| add offset data (ablation) | PointTrack · KITTI MOTS val (cars) | sMOTSA: 85.32 -> 85.51 | Improved |
| add category data (ablation) | PointTrack · KITTI MOTS val (cars) | sMOTSA: 85.33 -> 85.51 | Improved |
| add position data (ablation) | PointTrack · KITTI MOTS val (cars) | sMOTSA: 85.35 -> 85.51 | No clear effect |
| replace DeepSort with PointTrack on same segmentation | PointTrack · APOLLO MOTS val | sMOTSA: 64.69 -> 70.76 | Improved |
| replace DeepSort with PointTrack on same segmentation | PointTrack · APOLLO MOTS val | IDS: 1692 -> 292 | Improved |
| fine-tune embedding only on KITTI MOTS (PointTrack*) instead of per-dataset training | PointTrack · MOTSChallenge | sMOTSA: 58.09 -> 57.98 | No clear effect |
Closest papers by meaning
- Segment as Points for Efficient and Effective Online Multi-Object Tracking and Segmentation (TPAMI 2021 · similarity 0.96)
- MOTS: Multi-Object Tracking and Segmentation (CVPR 2019 · similarity 0.92)
- TrackIME: Enhanced Video Point Tracking via Instance Motion Estimation (NeurIPS 2024 · similarity 0.87)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.87)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.87)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · 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.
