Research map / Label-efficient and robust learning / Recognition and long-tail
Explainable Person Re-Identification With Attribute-Guided Metric Distillation
Xiaodong Chen, Xinchen Liu, Wu Liu, Xiao-Ping Zhang, Yongdong Zhang, Tao Mei
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 interpreter distance re-weighting (ablation) | SBS (R101) · DukeMTMC-ReID | Rank-1 / mAP: 89.3 -> 90.2; 78.4 -> 79.1 | Improved |
| add interpreter distance re-weighting (ablation) | SBS (R50) · DukeMTMC-ReID | Rank-1 / mAP: 88.2 -> 89.1; 75.5 -> 75.6 | Improved |
| add interpreter distance re-weighting (ablation) | BOT (R50) · Market-1501 | Rank-1 / mAP: 93.8 -> 94.4; 84.7 -> 86.1 | Improved |
| add interpreter distance re-weighting (ablation) | OSNet · Market-1501 | Rank-1 / mAP: 94.7 -> 95.0; 85.7 -> 86.1 | Improved |
| replace target model with attribute-guided interpreter | SBS (ResNet-50) · M->D cross-domain | Rank-1 / mAP: 45.02 -> 47.08; 26.43 -> 28.41 | Improved |
| replace target model with attribute-guided interpreter | SBS (ResNet-50) · D->M cross-domain | Rank-1 / mAP: 53.15 -> 54.48; 24.48 -> 25.48 | Improved |
| distill metric into interpreter (in-domain) | SBS (ResNet-101) · Market-1501 | Rank-1: 95.94 -> 95.55 | Hurt |
| distill metric into interpreter (in-domain) | SBS (ResNet-34) · Market-1501 | Rank-1: 93.94 -> 94.21 | No clear effect |
| share 3 stages instead of 1 with target model (ablation) | interpreter · Market-1501 | Rank-1: 94.76 -> 94.74 | No clear effect |
| share 5 stages instead of 3 (ablation) | interpreter · Market-1501 | Rank-1: 94.74 -> 94.89 | No clear effect |
| share 2 stages instead of 3 (ablation) | interpreter · Market-1501 | Rank-1: 94.74 -> 93.98 | Hurt |
Closest papers by meaning
- Multi-Prompts Learning with Cross-Modal Alignment for Attribute-Based Person Re-identification (AAAI 2024 · similarity 0.89)
- Deep Attributes Driven Multi-camera Person Re-identification (ECCV 2016 · similarity 0.88)
- Disentangled Representations for Short-Term and Long-Term Person Re-Identification (TPAMI 2021 · similarity 0.88)
- Semi-Supervised Bayesian Attribute Learning for Person Re-Identification (AAAI 2018 · similarity 0.88)
- Relation Network for Person Re-Identification (AAAI 2020 · similarity 0.88)
- AANet: Attribute Attention Network for Person Re-Identifications (CVPR 2019 · 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.
