Research map / Detection and segmentation / CNN features
Neural Person Search Machines
Hao Liu, Jiashi Feng, Zequn Jie, Karlekar Jayashree, Bo Zhao, Meibin Qi, Jianguo Jiang, Shuicheng Yan
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 attention prediction model on top of primitive memory (ablation) | NPSM w/o A&C · CUHK-SYSU (gallery 100, ablation) | mAP (%): 56.5 -> 72.5 | Improved |
| add attention prediction model on top of primitive memory (ablation) | NPSM w/o A&C · CUHK-SYSU (gallery 100, ablation) | top-1 (%): 62.5 -> 76.3 | Improved |
| add contextual cue memory mechanism (ablation) | NPSM w/o C · CUHK-SYSU (gallery 100, ablation) | mAP (%): 72.5 -> 77.9 | Improved |
| add contextual cue memory mechanism (ablation) | NPSM w/o C · CUHK-SYSU (gallery 100, ablation) | top-1 (%): 76.3 -> 81.2 | Improved |
| add recursive primitive-memory search without attention or context to OIM baseline (ablation) | OIM · CUHK-SYSU (gallery 100, ablation) | mAP (%): 75.5 -> 56.5 | Hurt |
| add recursive primitive-memory search without attention or context to OIM baseline (ablation) | OIM · CUHK-SYSU (gallery 100, ablation) | top-1 (%): 78.7 -> 62.5 | Hurt |
| add attention but no contextual cue to OIM baseline (ablation) | OIM · CUHK-SYSU (gallery 100, ablation) | mAP (%): 75.5 -> 72.5 | Hurt |
| replace OIM two-stage search with full NPSM | OIM · CUHK-SYSU (gallery 100) | mAP (%): 75.5 -> 77.9 | Improved |
| replace OIM two-stage search with full NPSM | OIM · CUHK-SYSU (gallery 100) | top-1 (%): 78.7 -> 81.2 | Improved |
| replace OIM with NPSM | OIM · PRW | mAP (%): 21.3 -> 24.2 | Improved |
| replace OIM with NPSM | OIM · PRW | top-1 (%): 49.9 -> 53.1 | Improved |
| replace OIM with NPSM across gallery sizes 50 to 4000 | OIM · CUHK-SYSU (varying gallery size) | mAP (%): around 2% | Improved |
Closest papers by meaning
- Efficient Person Search: An Anchor-Free Approach (IJCV 2023 · similarity 0.86)
- Divide and Conquer: Hybrid Pre-training for Person Search (AAAI 2024 · similarity 0.86)
- OIMNet++: Prototypical Normalization and Localization-Aware Learning for Person Search (ECCV 2022 · similarity 0.86)
- Diverse Knowledge Distillation for End-to-End Person Search (AAAI 2021 · similarity 0.86)
- Person Search With Natural Language Description (CVPR 2017 · similarity 0.85)
- Domain Generalizable Person Search Using Unreal Dataset (AAAI 2024 · 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.
