Research map / Video / Video segmentation and tracking
Visio-Temporal Attention for Multi-Camera Multi-Target Association
Yu-Jhe Li, Xinshuo Weng, Yan Xu, Kris M. Kitani
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 position embeddings (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 84.2 -> 85.5 | Improved |
| add position embeddings (ablation) | Ours · ConstructSite MTMC (GT tracklets) | IDF1: 90.21 -> 92.38 | Improved |
| keep temporal order of sampled images instead of shuffling (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 83.9 -> 85.5 | Improved |
| keep temporal order of sampled images instead of shuffling (ablation) | Ours · ConstructSite MTMC (GT tracklets) | IDF1: 90.37 -> 92.38 | Improved |
| add intra-tracklet attention Hintra (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 82.7 -> 85.5 | Improved |
| add intra-tracklet attention Hintra (ablation) | Ours · ConstructSite MTMC (GT tracklets) | IDF1: 88.77 -> 92.38 | Improved |
| add inter-tracklet attention Hinter (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 78.5 -> 85.5 | Improved |
| add inter-tracklet attention Hinter (ablation) | Ours · ConstructSite MTMC (GT tracklets) | IDF1: 80.49 -> 92.38 | Improved |
| reduce heads from 12 to 6 in both modules (ablation) | Ours · WILDTRACK (video Re-ID) | mAP: 72.0 -> 64.5 | Hurt |
| reduce heads to 1 in both modules (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 85.5 -> 75.7 | Hurt |
| double heads from 12 to 24 (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 85.5 -> 85.4 | No clear effect |
| increase sampling T from 1 to 8 (ablation) | Ours · WILDTRACK (video Re-ID) | Rank-1: 77.4 -> 85.5 | Improved |
Closest papers by meaning
- Unsupervised Graph Association for Person Re-Identification (ICCV 2019 · similarity 0.88)
- Self-supervised Multi-view Person Association and Its Applications (TPAMI 2020 · similarity 0.88)
- Tracklet Self-Supervised Learning for Unsupervised Person Re-Identification (AAAI 2020 · similarity 0.87)
- DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion (CVPR 2022 · similarity 0.87)
- Features for Multi-Target Multi-Camera Tracking and Re-Identification (CVPR 2018 · similarity 0.87)
- ADA-Track: End-to-End Multi-Camera 3D Multi-Object Tracking with Alternating Detection and Association (CVPR 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.
