Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Progressive Modality-Shared Transformers for Effective Visible-Infrared Person Re-identification
Read the paperCheck 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 the Gray-IR then RGB-IR progressive schedule instead of RGB-IR training (ablation) | PMT · SYSU-MM01 all-search | Rank-1: 52.51 -> 59.07 | Helped |
| use progressive hard triplet loss rather than hard triplet loss (ablation) | Gray-RGB progressive baseline · SYSU-MM01 all-search | Rank-1: 59.07 -> 61.67 | Helped |
| replace hard triplet loss with WRT (ablation) | Gray-RGB progressive baseline · SYSU-MM01 all-search | Rank-1: 59.07 -> 54.90 | Hurt |
| add MSEL with Euclidean distance (ablation) | Base (PL) · SYSU-MM01 all-search | Rank-1: 61.67 -> 65.11 | Helped |
| use Euclidean distance rather than cosine distance in MSEL (ablation) | PMT · SYSU-MM01 all-search | Rank-1: 64.19 -> 65.11 | Helped |
| add DCL with dynamic negative-sample selection (ablation) | Base (PL) · SYSU-MM01 all-search | mAP: 59.26 -> 62.76 | Helped |
| use dynamic negative selection rather than selecting all negatives in DCL (ablation) | DCL · SYSU-MM01 all-search | Rank-1: 62.86 -> 64.05 | Helped |
| combine Euclidean MSEL with dynamic DCL (ablation) | Base (PL) · SYSU-MM01 all-search | Rank-1: 65.11 -> 67.53 | Helped |
| add MSEL to the AGW CNN backbone (ablation) | AGW · SYSU-MM01 all-search | Rank-1: 58.19 -> 62.16 | Helped |
| add progressive learning and DCL to AGW+MSEL (ablation) | AGW CNN backbone · SYSU-MM01 all-search | Rank-1: 62.16 -> 67.09 | Helped |
| use PMT rather than the strongest reported baseline | SYSU-MM01 all-search · SYSU-MM01 all-search | Rank-1: 65.34 -> 67.53 | Helped |
| use PMT rather than the strongest reported baseline | PMT · RegDB V to T | Rank-1: 80.54 -> 84.83 | Helped |
Closest papers by meaning
- Cross-Modality Person Re-identification with Memory-Based Contrastive Embedding (AAAI 2023 · similarity 0.93)
- Empowering Visible-Infrared Person Re-Identification with Large Foundation Models (NeurIPS 2024 · similarity 0.93)
- Implicit Discriminative Knowledge Learning for Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.92)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.92)
- High-Order Structure Based Middle-Feature Learning for Visible-Infrared Person Re-identification (AAAI 2024 · similarity 0.89)
- Learning Commonality, Divergence and Variety for Unsupervised Visible-Infrared Person Re-identification (NeurIPS 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.