Research map / Detection and segmentation / CNN features
BPJDet: Extended Object Representation for Generic Body-Part Joint Detection
Huayi Zhou, Fei Jiang, Jiaxin Si, Yue Ding, Hongtao Lu
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 body-head association to double check head boxes (ablation) | BPJDet-S · CrowdHuman val head detection | precision: 69.7 -> 81.1 | Improved |
| add body-head association (ablation) | BPJDet-S · SCUT Head Part_B | precision: 84.5 -> 90.4 | Improved |
| add body-head association (ablation) | BPJDet-S · CroHD train-set | precision: 49.9 -> 61.5 | Improved |
| add body-head association (ablation) | BPJDet-S · CroHD train-set | mAP: 35.9 -> 34.7 | Hurt |
| add association task (extended representation) to detector training (ablation) | BPJDet · CityPersons | AP body and face | No clear effect |
| add association task to detector (ablation) | BPJDet · CityPersons | body-part mMR-2 | Improved |
| set loss weight lambda to 0.015 instead of larger or smaller (ablation) | BPJDet-S · CityPersons | MR-2 and mMR-2 | Improved |
| set inner IoU threshold to 0.6 within 0.4-0.8 (ablation) | BPJDet-S · BodyHands | Cond. Accuracy and Joint AP | Mixed |
| use anchor-free YOLOv5u instead of anchor-based YOLOv5 | BPJDet-L · CityPersons | average mMR-2: 31.5 -> 29.3 | Improved |
| use BPJDet-L5u instead of previous best BodyHands method | BPJDet · BodyHands | Joint AP: 63.87 -> 85.42 | Improved |
| use BPJDet-L8u instead of CrowdDet+BFJ | BPJDet · CrowdHuman body-face | mMR-2: 52.3 -> 49.9 | Improved |
| train body jointly with head instead of face | BPJDet-L5u · CrowdHuman | AP body: 91.8 -> 87.4 | Hurt |
Closest papers by meaning
- PBADet: A One-Stage Anchor-Free Approach for Part-Body Association (ICLR 2024 · similarity 0.90)
- Body-Face Joint Detection via Embedding and Head Hook (ICCV 2021 · similarity 0.90)
- PairDETR : Joint Detection and Association of Human Bodies and Faces (CVPR 2024 · similarity 0.89)
- Relational Learning for Joint Head and Human Detection (AAAI 2020 · similarity 0.88)
- GigaHumanDet: Exploring Full-Body Detection on Gigapixel-Level Images (AAAI 2024 · similarity 0.87)
- Detection Based Part-level Articulated Object Reconstruction from Single RGBD Image (NeurIPS 2023 · similarity 0.87)
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.
