Research map / Detection and segmentation / CNN features
Beta R-CNN: Looking into Pedestrian Detection from Another Perspective
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 BetaHead (regress 8 beta parameters) (ablation) | Cascade R-CNN · CrowdHuman val | MR-2: 43.8 -> 43.5 | Helped |
| add BetaHead (regress 8 beta parameters) (ablation) | Cascade R-CNN · CrowdHuman val | AP: 85.2 -> 85.5 | Helped |
| add BetaMask on top of BetaHead (ablation) | Beta R-CNN · CrowdHuman val | MR-2: 43.5 -> 41.3 | Helped |
| add BetaMask on top of BetaHead (ablation) | Beta R-CNN · CrowdHuman val | AP: 85.5 -> 87.1 | Helped |
| add Mask Loss (KL supervision of BetaMask) (ablation) | Beta R-CNN · CrowdHuman val | MR-2: 41.3 -> 41.1 | Helped |
| add BetaNMS (KL threshold 7) replacing IoU NMS (ablation) | Beta R-CNN · CrowdHuman val | MR-2: 41.1 -> 40.3 | Helped |
| add BetaNMS (KL threshold 7) replacing IoU NMS (ablation) | Beta R-CNN · CrowdHuman val | AP: 87.2 -> 88.2 | Helped |
| replace full-body IoU NMS with visible-box IoU NMS (ablation) | Beta R-CNN · CrowdHuman val | MR-2: 41.1 -> 42.0 | Hurt |
| replace full-body IoU NMS with visible-box IoU NMS (ablation) | Beta R-CNN · CrowdHuman val | AP: 87.2 -> 88.1 | Helped |
| combine fIoU and vIoU NMS vs fIoU NMS alone (ablation) | Beta R-CNN · CrowdHuman val | AP: 87.2 -> 88.1 | Helped |
| replace SoftNMS with BetaNMS (KL 7) (ablation) | Beta R-CNN · CrowdHuman val | MR-2: 41.1 -> 40.3 | Helped |
| lower BetaNMS KL threshold from 7 to 6 (ablation) | Beta R-CNN · CrowdHuman val | AP: 88.2 -> 89.6 | Mixed |
Closest papers by meaning
- Sparse R-CNN: End-to-End Object Detection With Learnable Proposals (CVPR 2021 · similarity 0.85)
- Where, What, Whether: Multi-Modal Learning Meets Pedestrian Detection (CVPR 2020 · similarity 0.84)
- NMS by Representative Region: Towards Crowded Pedestrian Detection by Proposal Pairing (CVPR 2020 · similarity 0.84)
- What Can Help Pedestrian Detection? (CVPR 2017 · similarity 0.84)
- High-Level Semantic Feature Detection: A New Perspective for Pedestrian Detection (CVPR 2019 · similarity 0.84)
- Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset (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.