Research map / Detection and segmentation / CNN features
RCAA: Relational Context-Aware Agents for Person Search
Xiaojun Chang, Po-Yao Huang, Yi-Dong Shen, Xiaodan Liang, Yi Yang, Alexander G. Hauptmann
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 relational spatial and temporal context (ablation) | RCAA · Person search benchmark | mAP (%): 69.8 -> 79.3 | Improved |
| add relational spatial context to temporal-only agent (ablation) | RCAA · Person search benchmark | mAP (%): 71.4 -> 79.3 | Improved |
| add relational spatial context to temporal-only agent (ablation) | RCAA · Person search benchmark | CMC top-1 (%): 74.8 -> 81.3 | Improved |
| use relational context feature instead of global context feature (ablation) | RCAA · Person search benchmark | mAP (%): 76.7 -> 79.3 | Improved |
| use relational context feature instead of global context feature (ablation) | RCAA · Person search benchmark | CMC top-1 (%): 78.9 -> 81.3 | Improved |
| add global context plus temporal context over no spatial context (ablation) | RCAA · Person search benchmark | mAP (%): 69.8 -> 76.7 | Improved |
| use RCAA instead of Joint Detection and Identification | person search · Person search benchmark (CNN detector) | mAP (%): 75.7 -> 79.3 | Improved |
| use RCAA instead of Joint Detection and Identification with ground-truth boxes | person search · Person search benchmark | mAP (%): 77.9 -> 79.3 | Improved |
| use off-the-shelf CCF detector instead of perfect GT detector | DSIFT + Euclidean · Person search benchmark | CMC top-1 (%): 45.9 -> 11.7 | Hurt |
| add more gallery images (increase gallery size from 50 to 6,978) | all compared methods · Person search benchmark | CMC top-1 and mAP: performance decreases as gallery size increases | Hurt |
Closest papers by meaning
- Exploring Visual Context for Weakly Supervised Person Search (AAAI 2022 · similarity 0.87)
- SCAN: A Spatial Context Attentive Network for Joint Multi-Agent Intent Prediction (AAAI 2021 · similarity 0.85)
- OIMNet++: Prototypical Normalization and Localization-Aware Learning for Person Search (ECCV 2022 · similarity 0.85)
- Learning Context Graph for Person Search (CVPR 2019 · similarity 0.85)
- A Survey on Text-Based Person Search (TPAMI 2026 · similarity 0.85)
- Optimizing Label Assignment for Weakly Supervised Person Search (AAAI 2025 · similarity 0.83)
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.
