Research map / Detection and segmentation / CNN features
PedHunter: Occlusion Robust Pedestrian Detector in Crowded Scenes
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 |
|---|---|---|---|
| add strict classification criterion (IoU 0.7 + jittered GT) (ablation) | FPN ResNet-50 · CityPersons | MR-2: 11.67 -> 10.59 | Helped |
| add strict classification criterion (ablation) | FPN ResNet-50 · CrowdHuman | MR-2: 46.8 -> 45.2 | Helped |
| add mask-guided head module (ablation) | FPN + strict · CityPersons | MR-2: 10.59 -> 9.62 | Helped |
| add mask-guided head module (ablation) | FPN + strict · CrowdHuman | MR-2: 45.2 -> 43.6 | Helped |
| add occlusion-simulated data augmentation (ablation) | FPN + strict + mask · CrowdHuman | MR-2: 43.6 -> 39.5 | Helped |
| add occlusion-simulated data augmentation (ablation) | FPN + strict + mask · CityPersons | MR-2: 9.62 -> 8.32 | Helped |
| all three components (ablation) | FPN baseline · CityPersons heavy occlusion | MR-2: 49.74 -> 43.53 | Helped |
| PedHunter vs OR-CNN | pedestrian detection · CityPersons val Reasonable | MR-2: 11.00 -> 8.32 | Helped |
Closest papers by meaning
- Improving Occlusion and Hard Negative Handling for Single-Stage Pedestrian Detectors (CVPR 2018 · similarity 0.88)
- Occluded Pedestrian Detection Through Guided Attention in CNNs (CVPR 2018 · similarity 0.87)
- NMS by Representative Region: Towards Crowded Pedestrian Detection by Proposal Pairing (CVPR 2020 · similarity 0.86)
- Temporal-Context Enhanced Detection of Heavily Occluded Pedestrians (CVPR 2020 · similarity 0.86)
- Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset (AAAI 2025 · similarity 0.85)
- Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting (CVPR 2025 · similarity 0.84)
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.