Research map / Label-efficient and robust learning / Recognition and long-tail
A Bottom-Up Clustering Approach to Unsupervised Person Re-Identification
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 diversity regularization to cluster merging (ablation) | BUC · Market-1501 | rank-1: 62.9 -> 66.2 | Helped |
| add diversity regularization to cluster merging (ablation) | BUC · DukeMTMC-reID | rank-1: 41.3 -> 47.4 | Helped |
| add diversity regularization to cluster merging (ablation) | BUC · MARS | rank-1: 55.5 -> 61.1 | Helped |
| add diversity regularization to cluster merging (ablation) | BUC · DukeMTMC-VideoReID | rank-1: 60.7 -> 69.2 | Helped |
| set diversity weight lambda above 0.05 (ablation) | BUC · Market-1501 | rank-1 | Hurt |
| use minimum-distance merging criterion instead of maximum (ablation) | BUC · Market-1501 | rank-1: 62.5 -> 66.2 | Helped |
| use minimum-distance merging criterion instead of centroid (ablation) | BUC · Market-1501 | rank-1: 65.8 -> 66.2 | No clear effect |
| iterative cluster merging (16 merges) (ablation) | BUC · Market-1501 | rank-1: 33.2% to 66.2% | Helped |
| BUC vs fully unsupervised OIM | re-ID · Market-1501 | rank-1: 38.0 -> 66.2 | Helped |
| BUC (no labels) vs one-example EUG | re-ID · MARS | rank-1: 62.6 -> 61.1 | Hurt |
Closest papers by meaning
- Unsupervised Person Re-Identification via Softened Similarity Learning (CVPR 2020 · similarity 0.92)
- Asymmetric Co-Teaching for Unsupervised Cross-Domain Person Re-Identification (AAAI 2020 · similarity 0.91)
- Camera-Aware Proxies for Unsupervised Person Re-Identification (AAAI 2021 · similarity 0.90)
- AD-Cluster: Augmented Discriminative Clustering for Domain Adaptive Person Re-Identification (CVPR 2020 · similarity 0.90)
- Divide-and-Regroup Clustering for Domain Adaptive Person Re-identification (AAAI 2022 · similarity 0.90)
- From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization (CVPR 2025 · similarity 0.88)
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.