Research map / 3D vision / Human pose estimation
RMPE: Regional Multi-Person Pose Estimation
Hao-Shu Fang, Shuqin Xie, Yu-Wing Tai, Cewu 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 symmetric STN (SSTN) and parallel SPPE (ablation) | RMPE · MPII multi-person validation | total mAP: 78.2 -> 80.8 | Improved |
| add parallel SPPE to SSTN (ablation) | RMPE · MPII multi-person validation | total mAP: 79.1 -> 80.8 | Improved |
| add pose-guided proposals generator (PGPG) (ablation) | RMPE · MPII multi-person validation | total mAP: 73.0 -> 80.8 | Improved |
| use PGPG instead of random jittering of boxes (ablation) | RMPE · MPII multi-person validation | total mAP: 77.9 -> 80.8 | Improved |
| add parametric pose NMS (ablation) | RMPE · MPII multi-person validation | total mAP: 75.7 -> 80.8 | Improved |
| use parametric pose NMS instead of prior pose NMS (variant 1) (ablation) | RMPE · MPII multi-person validation | total mAP: 79.1 -> 80.8 | Improved |
| use parametric pose NMS instead of prior pose NMS (variant 2) (ablation) | RMPE · MPII multi-person validation | total mAP: 79.9 -> 80.8 | Improved |
| use RMPE instead of straight forward two-step pipeline (ablation) | RMPE · MPII multi-person validation | total mAP: 71.7 -> 80.8 | Improved |
| replace detector proposals with ground truth boxes (ablation) | RMPE · MPII multi-person validation | total mAP: 80.8 -> 84.2 | Improved |
| use faster-rcnn with softnms and PyraNet instead of SSD-512 and hourglass | RMPE · MPII multi-person test | total mAP: 76.7 -> 82.1 | Improved |
| use parametric pose NMS rather than prior implementation (speed) | pose NMS · MPII multi-person validation | time on 1300 images: 62.2 seconds -> 1.8 seconds | Improved |
| use stronger detector and pose estimator on COCO | RMPE · MSCOCO keypoints test-dev | AP: 61.8 -> 72.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DeeperCut — DeeperCut: A Deeper, Stronger, and Faster Multi-person Pose Estimation Model (ECCV 2016) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- End-to-End Multi-Person Pose Estimation with Pose-Aware Video Transformer (AAAI 2026 · similarity 0.88)
- Dual Networks Based 3D Multi-Person Pose Estimation From Monocular Video (TPAMI 2022 · similarity 0.88)
- End-to-End Multi-Person Pose Estimation With Transformers (CVPR 2022 · similarity 0.88)
- AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time (TPAMI 2022 · similarity 0.88)
- PoseTrack: Joint Multi-Person Pose Estimation and Tracking (CVPR 2017 · similarity 0.87)
- Towards Accurate Multi-Person Pose Estimation in the Wild (CVPR 2017 · 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.
