Research map / Label-efficient and robust learning / Self-supervised and contrastive
Unveiling the Power of Self-Supervision for Multi-View Multi-Human Association and Tracking
Wei Feng, Feifan Wang, Ruize Han, Yiyang Gan, Zekun Qian, Junhui Hou, Song Wang
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 |
|---|---|---|---|
| self-supervised end-to-end method vs prior version (dummy nodes, spatial-temporal global info) | Prior version · MvMHAT | AF1: 49.5 -> 59.6 | Improved |
| same, vs prior version | Prior version · MMP-MvMHAT | AF1: 26.5 -> 50.4 | Improved |
| same, vs prior version | Prior version · MMP-MvMHAT | IDF1: 58.7 -> 67.3 | Improved |
| end-to-end training with consistency vs two-stage SVT | SVT · MvMHAT | F: 50.0 -> 58.8 | Improved |
| end-to-end training vs two-stage SVT | SVT · MMP-MvMHAT | F: 52.3 -> 58.9 | Improved |
| end-to-end vs two-stage DeepCC | DeepCC · MvMHAT | AF1: 43.4 -> 59.6 | Improved |
| multi-view method vs single-view MOT TrackFormer | TrackFormer · MvMHAT | AF1: 31.3 -> 59.6 | Improved |
| use private detector for TraDeS | TraDeS · MvMHAT | IDF1: 44.9 -> 52.1 | Improved |
| use private detector for CenterTrack | CenterTrack · MvMHAT | IDF1: 38.1 -> 38.1 | No clear effect |
| our method vs SVT on MOTA (over-time detection accuracy) | SVT · MMP-MvMHAT | MOTA: 96.7 -> 95.0 | Hurt |
| our method vs TraDeS (P) on MOTA | TraDeS (P) · MvMHAT | MOTA: 70.8 -> 66.3 | Hurt |
Closest papers by meaning
- Self-supervised Multi-view Person Association and Its Applications (TPAMI 2020 · similarity 0.90)
- Learning from Synchronization: Self-Supervised Uncalibrated Multi-View Person Association in Challenging Scenes (CVPR 2025 · similarity 0.89)
- Self-Supervised Multi-Object Tracking with Cross-input Consistency (NeurIPS 2021 · similarity 0.87)
- Self-Supervised Multi-Object Tracking with Path Consistency (CVPR 2024 · similarity 0.87)
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.85)
- CoMotion: Concurrent Multi-person 3D Motion (ICLR 2025 · similarity 0.85)
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.
