Research map / Detection and segmentation / CNN features
NMS by Representative Region: Towards Crowded Pedestrian Detection by Proposal Pairing
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 |
|---|---|---|---|
| Use paired full/visible boxes with mask PPFE and R²NMS instead of the reimplemented baseline | FPN ResNet-50 pedestrian detector · CrowdHuman val | MR: 46.28 -> 43.35 | Helped |
| Use paired full/visible boxes with mask PPFE and R²NMS instead of the reimplemented baseline | FPN ResNet-50 pedestrian detector · CrowdHuman val | AP: 84.91 -> 89.29 | Helped |
| Use paired full/visible boxes with mask PPFE and R²NMS instead of the reimplemented baseline | FPN ResNet-50 pedestrian detector · CrowdHuman val | Recall: 88.25 -> 93.33 | Helped |
| Replace standard NMS with R²NMS in NPM (ablation) | NPM · CrowdHuman val | MR: 45.30 -> 45.87 | Hurt |
| Replace standard NMS with R²NMS in NPM (ablation) | NPM · CrowdHuman val | AP: 85.59 -> 88.39 | Helped |
| Replace standard NMS with R²NMS in NPM (ablation) | NPM · CrowdHuman val | Recall: 88.69 -> 93.76 | Helped |
| Replace standard NMS with R²NMS in PBM (ablation) | PBM with concat PPFE · CrowdHuman val | MR: 44.79 -> 43.57 | Helped |
| Replace standard NMS with R²NMS in PBM (ablation) | PBM with concat PPFE · CrowdHuman val | AP: 85.62 -> 89.28 | Helped |
| Add concatenation-based PPFE to NPM (ablation) | Paired-box detector · CrowdHuman val | MR: 45.43 -> 44.32 | Helped |
| Add concatenation-based PPFE to NPM (ablation) | Paired-box detector · CrowdHuman val | MR-V: 54.18 -> 52.19 | Helped |
| Add mask-attention PPFE to NPM (ablation) | Paired-box detector · CrowdHuman val | MR: 45.43 -> 44.20 | Helped |
| Use mask-attention rather than concatenation PPFE (ablation) | PBM with standard NMS · CrowdHuman val | MR: 44.32 -> 44.20 | No clear effect |
Closest papers by meaning
- Adaptive NMS: Refining Pedestrian Detection in a Crowd (CVPR 2019 · similarity 0.90)
- Detection in Crowded Scenes: One Proposal, Multiple Predictions (CVPR 2020 · similarity 0.88)
- PedHunter: Occlusion Robust Pedestrian Detector in Crowded Scenes (AAAI 2020 · similarity 0.86)
- Repulsion Loss: Detecting Pedestrians in a Crowd (CVPR 2018 · similarity 0.86)
- Accelerating Non-Maximum Suppression: A Graph Theory Perspective (NeurIPS 2024 · similarity 0.84)
- Union-over-Intersections: Object Detection beyond Winner-Takes-All (ICLR 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.