Research map / Label-efficient and robust learning / Recognition and long-tail
Unsupervised Visible-Infrared Person Re-identification under Unpaired Settings
Haoyu Yao, Bin Yang, Wenke Huang, Bo Du, Mang Ye
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 CFM and static learning (multi-memory banks) (ablation) | ADCA baseline · SYSU-MM01 unpaired all-search alpha=0.5 | mAP: 47.29 -> 52.48 (5.19% mAP gain) | Improved |
| add CFM and static learning (multi-memory banks) (ablation) | ADCA baseline · RegDB unpaired visible-to-infrared alpha=0.5 | mAP: 48.6 -> 54.52 (5.92% mAP gain) | Improved |
| add CFM and dynamic learning (CLP loss) (ablation) | ADCA baseline · SYSU-MM01 unpaired all-search alpha=0.5 | mAP: 47.29 -> 50.67 (3.38% mAP gain) | Improved |
| add CFM and dynamic learning (CLP loss) (ablation) | ADCA baseline · RegDB unpaired visible-to-infrared alpha=0.5 | mAP: 48.6 -> 50.84 (2.24% mAP gain) | Improved |
| add dynamic learning on top of static learning (collaborative) (ablation) | MCL · SYSU-MM01 unpaired all-search alpha=0.5 | mAP: 52.48 -> 53.98 | Improved |
| add dynamic learning on top of static learning (collaborative) (ablation) | MCL · RegDB unpaired visible-to-infrared alpha=0.5 | mAP: 54.52 -> 59.69 | Improved |
| add static learning on top of dynamic learning (collaborative) (ablation) | MCL · SYSU-MM01 unpaired all-search alpha=0.5 | mAP: 50.67 -> 53.98 | Improved |
| increase unpaired ratio alpha from 0.25 to 1.0 | ADCA · SYSU-MM01 unpaired all-search | rank-1: 41.80 -> 23.48 | Hurt |
| increase unpaired ratio alpha from 0.25 to 1.0 | MCL · SYSU-MM01 unpaired all-search | rank-1: 58.57 -> 45.98 | Hurt |
| replace N-ULC with MCL at fully unpaired alpha=1.0 | MCL vs N-ULC · SYSU-MM01 unpaired all-search alpha=1.0 | rank-1: 34.40 -> 45.98 | Improved |
| replace N-ULC with MCL at fully unpaired alpha=1.0 | MCL vs N-ULC · RegDB unpaired visible-to-infrared alpha=1.0 | rank-1: 34.71 -> 43.18 | Improved |
| apply MCL under paired setting instead of ADCA | MCL vs ADCA · SYSU-MM01 paired all-search | rank-1: 45.51 -> 62.95 | Improved |
Closest papers by meaning
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.94)
- Multi-Memory Matching for Unsupervised Visible-Infrared Person Re-Identification (ECCV 2024 · similarity 0.94)
- Modality Unifying Network for Visible-Infrared Person Re-Identification (ICCV 2023 · similarity 0.94)
- Empowering Visible-Infrared Person Re-Identification with Large Foundation Models (NeurIPS 2024 · similarity 0.92)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.92)
- Implicit Discriminative Knowledge Learning for Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.92)
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.
