Research map / Label-efficient and robust learning / Recognition and long-tail
Intra-Inter Camera Similarity for Unsupervised Person Re-Identification
Check 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 intra-camera pseudo-label training before cross-camera training (ablation) | IICS · Market1501 | Rank-1: 48.8 -> 78.6 | Helped |
| add intra-camera pseudo-label training before cross-camera training (ablation) | IICS · DukeMTMC-ReID | Rank-1: 16.7 -> 54.2 | Helped |
| add inter-camera training using raw CNN feature similarity (ablation) | intra-camera-trained IICS · DukeMTMC-ReID | Rank-1: 62.9 -> 54.2 | Hurt |
| use the proposed inter-camera similarity instead of CNN feature similarity (ablation) | Stage1 + Stage 2 · Market1501 | Rank-1: 78.6 -> 88.8 | Helped |
| use the proposed inter-camera similarity instead of CNN feature similarity (ablation) | Stage1 + Stage 2 · DukeMTMC-ReID | Rank-1: 54.2 -> 76.9 | Helped |
| combine IN and BN statistics with AIBN (ablation) | ResNet-50 · Market1501 | Rank-1: 85.5 -> 88.8 | Helped |
| combine IN and BN statistics with AIBN (ablation) | ResNet-50 · DukeMTMC-ReID | Rank-1: 71.3 -> 76.9 | Helped |
| replace BN with IN alone (ablation) | ResNet-50 · Market1501 and DukeMTMC-ReID | Rank-1: 85.5 -> 83.2; 71.3 -> 72.7 | Mixed |
| replace BN with IBN (ablation) | ResNet-50 · Market1501 and DukeMTMC-ReID | Rank-1: 85.5 -> 81.1; 71.3 -> 56.3 | Hurt |
| optimize AIBN mixture weight rather than fixing it at 0.5 (ablation) | AIBN · Market1501 and DukeMTMC-ReID | Rank-1: 88.0 -> 88.8; 75.2 -> 76.9 | Helped |
| add AIBN to the cross-dataset transfer model (ablation) | ResNet-50 transfer model · DukeMTMC-ReID to Market1501 | Rank-1: 48.9 -> 54.5 | Helped |
| add AIBN to the cross-dataset transfer model (ablation) | ResNet-50 transfer model · Market1501 to DukeMTMC-ReID | Rank-1: 38.2 -> 49.5 | Helped |
Closest papers by meaning
- Camera-Aware Proxies for Unsupervised Person Re-Identification (AAAI 2021 · similarity 0.92)
- Unsupervised Person Re-Identification by Soft Multilabel Learning (CVPR 2019 · similarity 0.91)
- Unsupervised Person Re-Identification via Softened Similarity Learning (CVPR 2020 · similarity 0.90)
- Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification (ICLR 2020 · similarity 0.89)
- Joint Noise-Tolerant Learning and Meta Camera Shift Adaptation for Unsupervised Person Re-Identification (CVPR 2021 · similarity 0.89)
- CDE-Learning: Camera Deviation Elimination Learning for Unsupervised Person Re-identification (AAAI 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.