Research map / Label-efficient and robust learning / Recognition and long-tail
Unsupervised Visible-Infrared Person Re-Identification via Progressive Graph Matching and Alternate Learning
Zesen Wu, 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 PGM and ACCL to DCL baseline (ablation) | unsupervised VI-ReID · SYSU-MM01 All search | r1(%): 39.98 -> 57.27 | Improved |
| replace PGM with bidirectional combined loss vs alternate Laccl (ablation) | ACCL · SYSU-MM01 All search | r1(%): 51.12 -> 57.27 | Improved |
| alternate unidirectional cross contrastive learning (Laccl) vs combined bidirectional Lccl (ablation) | BGM setting · SYSU-MM01 All search | r1(%): 45.52 -> 51.98 | Improved |
| progressive graph matching instead of basic graph matching (ablation) | cross-modality correspondence mining · SYSU-MM01 All search | r1(%): 48.82 -> 51.98 (BGM vs PGM with Laccl, rows 2 vs 3 style) | Improved |
| progressive graph matching instead of direct matching (DM) (ablation) | correspondence accuracy · SYSU-MM01 All search | matching accuracy | Improved |
| add PGM+ACCL to DCL baseline (ablation) | unsupervised VI-ReID · RegDB Visible to Thermal | r1(%): 43.78 -> 69.48 | Improved |
| set cross-modality weight mu to 0 (no ACCL) (ablation) | ACCL loss weight · SYSU-MM01 All search | r1(%): 57.27 at mu=0.5 -> 39.98 at mu=0 (approx. DCL-only) | Hurt |
| vary mu in {0.3,0.5,0.7,1.0} (ablation) | ACCL loss weight · SYSU-MM01 All search | r1(%): 55.03 to 57.27 across mu in 0.3-1.0 range | No clear effect |
Closest papers by meaning
- Dual-Level Matching With Outlier Filtering for Unsupervised Visible-Infrared Person Re-Identification (TPAMI 2025 · similarity 0.96)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.92)
- Multi-Memory Matching for Unsupervised Visible-Infrared Person Re-Identification (ECCV 2024 · similarity 0.92)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.90)
- Learning Commonality, Divergence and Variety for Unsupervised Visible-Infrared Person Re-identification (NeurIPS 2024 · similarity 0.89)
- Implicit Discriminative Knowledge Learning for Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.89)
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.
