Research map / Detection and segmentation / CNN features
Explicit Box Detection Unifies End-to-End Multi-Person Pose Estimation
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 explicit human detection supervision (ablation) | ED-Pose · COCO val2017 | AP: +4.5 AP | Helped |
| add explicit human detection supervision (ablation) | ED-Pose · CrowdPose test | AP: +9.9 AP | Helped |
| represent keypoints with box width and height instead of only 2D coordinates (ablation) | ED-Pose · COCO val2017 | AP: 58.3 -> 71.6 | Helped |
| initialize keypoint box sizes with learnable embeddings weighted by the human box (ablation) | ED-Pose · COCO val2017 | AP: 70.8 -> 71.6 | Helped |
| allow external keypoint-keypoint attention in full query interaction (ablation) | ED-Pose · COCO val2017 | AP: 71.6 -> 70.4 | Hurt |
| add internal human-to-keypoint interaction (ablation) | ED-Pose interactive decoder · COCO val2017 | AP: 71.2 -> 71.6 | Helped |
| add external human-to-human interaction (ablation) | ED-Pose interactive decoder · COCO val2017 | AP: 71.3 -> 71.6 | Helped |
| increase selected human queries from 50 to 100 (ablation) | ED-Pose · COCO val2017 | AP: 70.9 -> 71.6 | Helped |
| increase selected human queries from 100 to 200 (ablation) | ED-Pose · COCO val2017 | AP: 71.6 -> 71.6 | No clear effect |
| use ED-Pose instead of PETR with Swin-L | multi-person pose estimation · COCO test-dev | AP: 70.5 -> 72.7 | Helped |
| use ED-Pose instead of HRFormer-B | multi-person pose estimation · CrowdPose test | AP: 72.4 -> 76.6 | Helped |
| use ED-Pose instead of PETR with ResNet-50 | multi-person pose estimation · COCO val2017 | inference time: 105 -> 51 | Helped |
Closest papers by meaning
- Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation (AAAI 2020 · similarity 0.90)
- Towards Accurate Multi-Person Pose Estimation in the Wild (CVPR 2017 · similarity 0.89)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.88)
- End-to-End Multi-Person Pose Estimation with Pose-Aware Video Transformer (AAAI 2026 · similarity 0.87)
- Bottom-Up Human Pose Estimation via Disentangled Keypoint Regression (CVPR 2021 · similarity 0.87)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · 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.