Research map / Label-efficient and robust learning / Recognition and long-tail
Weakly Supervised Visible-Infrared Person Re-Identification via Heterogeneous Expert Collaborative Consistency Learning
Yafei Zhang, Lingqi Kong, Huafeng Li, Jie Wen
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 cross-modal consistency learning (CMCL) with expert-consistent predictions (ablation) | Baseline (HEL) · SYSU-MM01 all search | Rank-1: 47.8 -> 66.7 | Improved |
| add CMCL (ablation) | Baseline (HEL) · SYSU-MM01 all search | mAP: 47.2 -> 62.8 | Improved |
| add CMCL (ablation) | Baseline (HEL) · SYSU-MM01 indoor search | Rank-1: 60.8 -> 73.4 | Improved |
| add cross-modal relationship establishment (CRE) fusing consistent and inconsistent predictions (ablation) | B+CMCL · SYSU-MM01 all search | Rank-1: 66.7 -> 68.0 | Improved |
| add CRE (ablation) | B+CMCL · SYSU-MM01 indoor search | mAP: 77.6 -> 78.3 | Improved |
| add collaborative learning among experts (CLAE) (ablation) | B+CRE+CMCL · SYSU-MM01 all search | Rank-1: 68.0 -> 70.4 | Improved |
| add CLAE (ablation) | B+CRE+CMCL · SYSU-MM01 indoor search | mAP: 78.3 -> 80.2 | Improved |
| use visible-expert predictions on infrared samples only (M r->v) instead of consistent matches (ablation) | B+CMCL · SYSU-MM01 all search | Rank-1: 66.7 -> 61.0 | Hurt |
| use infrared-expert predictions on visible samples only (M v->r) instead of consistent matches (ablation) | B+CMCL · SYSU-MM01 all search | Rank-1: 66.7 -> 67.7 | No clear effect |
| full CRE fusion vs infrared-expert-only relations (ablation) | B+CMCL · SYSU-MM01 all search | mAP: 64.4 -> 64.6 | Improved |
Closest papers by meaning
- Infrared-Visible Cross-Modal Person Re-Identification with an X Modality (AAAI 2020 · similarity 0.93)
- Learning by Aligning: Visible-Infrared Person Re-Identification Using Cross-Modal Correspondences (ICCV 2021 · similarity 0.92)
- Syncretic Modality Collaborative Learning for Visible Infrared Person Re-Identification (ICCV 2021 · similarity 0.92)
- Unsupervised Visible-Infrared Person Re-identification under Unpaired Settings (ICCV 2025 · similarity 0.92)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.91)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.90)
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.
