Research map / Label-efficient and robust learning / Recognition and long-tail
Hierarchical Discriminative Learning for Visible Thermal 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 |
|---|---|---|---|
| HCML metric learning instead of Euclidean distance (ablation) | TONE 2 features · RegDB visible-to-thermal | rank-1: 16.87 -> 24.44 | Helped |
| HCML instead of XQDA | cross-modal matching · RegDB visible-to-thermal | mAP: 21.80 -> 20.80 | Hurt |
| add modality-specific metric terms (HCML vs CML) (ablation) | metric learning · RegDB visible-to-thermal | rank-1 and mAP | Helped |
| partially shared two-stream network instead of one-stream (ablation) | TONE · RegDB visible-to-thermal | rank-1: 13.11 -> 16.87 | Helped |
| partially shared two-stream network instead of fully separate streams (ablation) | TONE · RegDB visible-to-thermal | rank-1: 1.43 -> 16.87 | Helped |
| add contrastive loss with alpha 0.2 to identity loss (ablation) | TONE · RegDB visible-to-thermal | rank-1: about 3% | Helped |
| use fc2 (TONE 2) instead of fc1 (TONE 1) features (ablation) | TONE with Euclidean distance · RegDB visible-to-thermal | rank-1: 11.07 -> 16.87 | Helped |
| add metric learning to concatenated modality features | modality fusion re-id · RegDB modality fusion | rank-1: 97.86 -> 98.72 | Helped |
| thermal query instead of visible query | HCML · RegDB | rank-1: 24.44 -> 21.70 | Hurt |
Closest papers by meaning
- Cross-Modality Person Re-identification with Memory-Based Contrastive Embedding (AAAI 2023 · similarity 0.93)
- Hi-CMD: Hierarchical Cross-Modality Disentanglement for Visible-Infrared Person Re-Identification (CVPR 2020 · similarity 0.92)
- Implicit Discriminative Knowledge Learning for Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.91)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (CVPR 2024 · similarity 0.90)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.89)
- Empowering Visible-Infrared Person Re-Identification with Large Foundation Models (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.