Research map / Label-efficient and robust learning / Domain adaptation and generalization
Unsupervised Person Re-identification by Deep Learning Tracklet Association
Minxian Li, Xiatian Zhu, Shaogang Gong
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 |
|---|---|---|---|
| use per-camera tracklet discrimination instead of joint-camera classification (ablation) | TAUDL without CCTA · CUHK03 | Rank-1 (%): 29.8 -> 33.8 | Improved |
| use per-camera tracklet discrimination instead of joint-camera classification (ablation) | TAUDL without CCTA · Market-1501 | Rank-1 (%): 17.5 -> 52.1 | Improved |
| use per-camera tracklet discrimination instead of joint-camera classification (ablation) | TAUDL without CCTA · DukeMTMC-reID | Rank-1 (%): 14.9 -> 51.2 | Improved |
| use per-camera tracklet discrimination instead of joint-camera classification (ablation) | TAUDL without CCTA · MARS | Rank-1 (%): 18.1 -> 38.0 | Improved |
| add cross-camera tracklet association (ablation) | TAUDL · CUHK03 | Rank-1 (%): 33.8 -> 44.7 | Improved |
| add cross-camera tracklet association (ablation) | TAUDL · Market-1501 | Rank-1 (%): 52.1 -> 63.7 | Improved |
| add cross-camera tracklet association (ablation) | TAUDL · DukeMTMC-reID | Rank-1 (%): 51.2 -> 61.7 | Improved |
| add cross-camera tracklet association (ablation) | TAUDL · MARS | Rank-1 (%): 38.0 -> 43.8 | Improved |
| increase ID duplication rate in tracklet labels from 0 to 50% (ablation) | TAUDL · MARS | Rank-1 (%): 43.8 -> 40.7 | Hurt |
| train on small video datasets (few tracklets) instead of MARS | TAUDL · PRID2011 / iLIDS-VID vs MARS | Rank-1 (%): MARS 43.8 vs 36.8 (best competitor); iLIDS-VID 26.7 vs 41.7 | Mixed |
Closest papers by meaning
- Unsupervised Tracklet Person Re-Identification (TPAMI 2019 · similarity 0.97)
- Tracklet Self-Supervised Learning for Unsupervised Person Re-Identification (AAAI 2020 · similarity 0.94)
- Stepwise Metric Promotion for Unsupervised Video Person Re-Identification (ICCV 2017 · similarity 0.92)
- Exploit the Unknown Gradually: One-Shot Video-Based Person Re-Identification by Stepwise Learning (CVPR 2018 · similarity 0.92)
- Weakly Supervised Tracklet Association Learning With Video Labels for Person Re-Identification (TPAMI 2023 · similarity 0.92)
- Unsupervised Person Re-Identification by Soft Multilabel Learning (CVPR 2019 · similarity 0.91)
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.
