Research map / Label-efficient and robust learning / Noisy labels
Learning With Twin Noisy Labels for Visible-Infrared 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 |
|---|---|---|---|
| Use DART to train with estimated annotation confidence and rectified cross-modal pairs instead of ADP training on noisy labels | ADP · SYSU-MM01, 20% and 50% noise; all-search and indoor-search | mAP: 23.71 -> 64.13; 10.83 -> 58.69; 34.97 -> 75.94; 20.81 -> 71.77 | Helped |
| Use DART to train with estimated annotation confidence and rectified cross-modal pairs instead of ADP training on noisy labels | ADP · RegDB, 20% and 50% noise; visible-to-infrared and infrared-to-visible | mAP: 35.92 -> 74.18; 34.75 -> 71.72; 11.25 -> 67.04; 12.31 -> 64.38 | Helped |
| Use DART rather than ADP-C trained after discarding noisy samples | ADP · SYSU-MM01, 20% and 50% noise; all-search and indoor-search | mAP: 61.57 -> 64.13; 56.49 -> 58.69; 73.82 -> 75.94; 69.05 -> 71.77 | Helped |
| Use DART rather than ADP-C trained after discarding noisy samples | ADP · RegDB, 20% and 50% noise; visible-to-infrared and infrared-to-visible | mAP: 70.02 -> 74.18; 68.95 -> 71.72; 66.75 -> 67.04; 63.05 -> 64.38 | Helped |
| Add co-modeling and soft identification loss (ablation) | ADP baseline (B) · SYSU-MM01 all-search, 20% noise | Rank-1: 25.44 -> 49.24 | Helped |
| Add pair division and special triplet loss (ablation) | ADP with co-modeling and soft identification loss · SYSU-MM01 all-search, 20% noise | Rank-1: 49.24 -> 65.44 | Helped |
| Add the quadruplet term for TP-FN and FP-TN triplets (ablation) | ADP with co-modeling, soft identification loss, and special triplet loss · SYSU-MM01 all-search, 20% noise | Rank-1: 65.44 -> 66.31 | Helped |
Closest papers by meaning
- Noisy-Correspondence Learning for Text-to-Image Person Re-identification (CVPR 2024 · similarity 0.89)
- Relieving Universal Label Noise for Unsupervised Visible-Infrared Person Re-Identification by Inferring from Neighbors (AAAI 2025 · similarity 0.88)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.87)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.86)
- High-Order Structure Based Middle-Feature Learning for Visible-Infrared Person Re-identification (AAAI 2024 · similarity 0.86)
- Learning Commonality, Divergence and Variety for Unsupervised Visible-Infrared Person Re-identification (NeurIPS 2024 · similarity 0.86)
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.