Research map / Detection and segmentation / CNN features
Instance Guided Proposal Network for Person Search
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 |
|---|---|---|---|
| Keep fewer proposals ranked by query similarity before person Re-id | IGPN + ResNet50 on CUHK-SYSU · CUHK-SYSU, 100 gallery size | mAP: 81.2 -> 85.3 | Helped |
| Keep fewer proposals ranked by query similarity before person Re-id | IGPN + PCB on CUHK-SYSU · CUHK-SYSU, 100 gallery size | mAP: 89.1 -> 90.3 | Helped |
| Use improved cross-correlation layers instead of vanilla layers (ablation) | IGPN · CUHK-SYSU, 100 gallery size | mAP: 84.2 -> 84.5 | No clear effect |
| Remove the shared Bsim/reg term from ICCL (ablation) | IGPN · CUHK-SYSU, 100 gallery size | mAP: 84.5 -> 84.0 | Hurt |
| Add the local relation block between proposals (ablation) | IGPN · CUHK-SYSU, 100 gallery size | mAP: 80.8 -> 84.5 | Helped |
| Add the global relation branch and its auxiliary search loss (ablation) | IGPN · CUHK-SYSU, 100 gallery size | mAP: 81.3 -> 84.5 | Helped |
| Add both local and global relation components (ablation) | IGPN · CUHK-SYSU, 100 gallery size | mAP: 79.2 -> 84.5 | Helped |
| Reduce proposals while retaining enough candidates for matching | IGPN + ResNet50 on PRW · PRW | mAP: 41.1 -> 42.9 | Helped |
| Reduce proposals while retaining enough candidates for matching | IGPN + PCB on PRW · PRW | mAP: 46.2 -> 47.2 | Helped |
| Reduce the number of proposals too far | IGPN + person Re-id on CUHK-SYSU · CUHK-SYSU, 100 gallery size | mAP: null | Hurt |
Closest papers by meaning
- Query-Guided End-To-End Person Search (CVPR 2019 · similarity 0.87)
- Sequential End-to-end Network for Efficient Person Search (AAAI 2021 · similarity 0.86)
- Diverse Knowledge Distillation for End-to-End Person Search (AAAI 2021 · similarity 0.86)
- Hierarchical Online Instance Matching for Person Search (AAAI 2020 · similarity 0.85)
- Divide and Conquer: Hybrid Pre-training for Person Search (AAAI 2024 · similarity 0.84)
- Cheb-GR: Rethinking K-nearest Neighbor Search in Re-ranking for Person Re-identification (CVPR 2025 · similarity 0.82)
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.