Research map / Vision-language and multimodal / CLIP and zero-shot
CLIP-ReID: Exploiting Vision-Language Model for Image Re-identification without Concrete Text Labels
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 two-stage rather than one-stage token and image-encoder training (ablation) | CLIP-ReID · MSMT17 | mAP: 68.9 -> 73.4 | Helped |
| Use two-stage rather than one-stage token and image-encoder training (ablation) | CLIP-ReID · MSMT17 | mAP: 61.9 -> 63.0 | Helped |
| Compare text embeddings against all training IDs rather than only batch IDs (ablation) | CLIP-ReID second stage · MSMT17 | mAP: 71.7 -> 73.4 | Helped |
| Remove the text-to-image loss from second-stage training (ablation) | CLIP-ReID second stage · MSMT17 | mAP: 73.2 -> 73.4 | Helped |
| Add triplet loss at the previous feature layer (ablation) | CLIP-ReID · MSMT17 | mAP: 73.1 -> 73.4 | Helped |
| Add SIE only to the CLS token (ablation) | CLIP-ReID · MSMT17 | mAP: 73.4 -> 74.5 | Helped |
| Add SIE to all global tokens instead of only CLS (ablation) | CLIP-ReID · MSMT17 | R1: 88.8 -> 88.6 | Hurt |
| Add overlapping patches (ablation) | CLIP-ReID · MSMT17 | mAP: 73.4 -> 74.6 | Helped |
| Add SIE to CLS and overlapping patches together (ablation) | CLIP-ReID · MSMT17 | mAP: 73.4 -> 75.8 | Helped |
| Use 4 learnable text tokens rather than 1 (ablation) | CLIP-ReID · MSMT17 | mAP: 72.7 -> 73.4 | Helped |
| Increase learnable text tokens from 4 to 8 (ablation) | CLIP-ReID · MSMT17 | mAP: 73.4 -> 73.4 | No clear effect |
| Use instance image features rather than averaged ID features in first-stage training (ablation) | CLIP-ReID · MSMT17 | mAP: 72.4 -> 73.4 | Helped |
Closest papers by meaning
- TF-CLIP: Learning Text-Free CLIP for Video-Based Person Re-identification (AAAI 2024 · similarity 0.88)
- Distilling CLIP with Dual Guidance for Learning Discriminative Human Body Shape Representation (CVPR 2024 · similarity 0.88)
- Instruct-ReID: A Multi-purpose Person Re-identification Task with Instructions (CVPR 2024 · similarity 0.87)
- CLIP-driven View-aware Prompt Learning for Unsupervised Vehicle Re-identification (AAAI 2025 · similarity 0.87)
- CLIMB-ReID: A Hybrid CLIP-Mamba Framework for Person Re-Identification (AAAI 2025 · similarity 0.87)
- SmartCLIP: Modular Vision-language Alignment with Identification Guarantees (CVPR 2025 · 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.