Research map / Vision-language and multimodal / Multimodal LLMs
ChatReID: Open-ended Interactive Person Retrieval via Hierarchical Progressive Tuning for Vision Language Models
Ke Niu, Haiyang Yu, Mengyang Zhao, Teng Fu, Siyang Yi, Wei Lu, Bin Li, Xuelin Qian, Xiangyang Xue
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 stage 2 fine-grained image retrieval tuning (ablation) | ChatReID · ten Re-ID benchmarks | average mAP: 41.0 -> 73.5 | Improved |
| add stage 1 person attribute understanding (ablation) | ChatReID · ten Re-ID benchmarks | average mAP: 73.5 -> 83.0 | Improved |
| add stage 1 to stage 3 without stage 2 (ablation) | ChatReID · ten Re-ID benchmarks | convergence | Hurt |
| scale backbone from 0.5B to 2B (ablation) | ChatReID · MSMT17 | Rank-1: 38.1 -> 90.1 | Improved |
| scale backbone from 0.5B to 2B (ablation) | ChatReID · RegDB | Rank-1: 4.3 -> 96.5 | Improved |
| scale backbone from 2B to 7B (ablation) | ChatReID · MSMT17 | Rank-1: 90.1 -> 91.3 | Improved |
| scale backbone from 2B to 7B (ablation) | ChatReID · CUHK-PEDES | Rank-1: 83.8 -> 87.7 | Improved |
| scale backbone from 2B to 7B (ablation) | ChatReID · LTCC | Rank-1: 82.7 -> 82.5 | No clear effect |
| scale backbone from 2B to 7B (ablation) | ChatReID · RegDB | Rank-1: 96.5 -> 96.3 | No clear effect |
| use ChatReID instead of IRM | standard Re-ID · MSMT17 | mAP: 72.4 -> 87.5 | Improved |
| use ChatReID instead of IRM | cloth-changing Re-ID · LTCC | mAP: 52.0 -> 76.3 | Improved |
| use ChatReID instead of DEEN | VI-ReID · SYSU-MM01 | mAP: 71.8 -> 83.6 | Improved |
Closest papers by meaning
- Human-centered Interactive Learning via MLLMs for Text-to-Image Person Re-identification (CVPR 2025 · similarity 0.90)
- Vision-Language Attribute Disentanglement and Reinforcement for Lifelong Person Re-Identification (CVPR 2026 · similarity 0.90)
- LLaVA-ReID: Selective Multi-image Questioner for Interactive Person Re-Identification (ICML 2025 · similarity 0.88)
- FlexiReID: Adaptive Mixture of Expert for Multi-Modal Person Re-Identification (ICML 2025 · similarity 0.88)
- Instruct-ReID: A Multi-purpose Person Re-identification Task with Instructions (CVPR 2024 · similarity 0.88)
- Alice Benchmarks: Connecting Real World Re-Identification with the Synthetic (ICLR 2024 · similarity 0.88)
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.
