Research map / 3D vision / Human pose estimation
Pose Partition Networks for Multi-Person Pose Estimation
Xuecheng Nie, Jiashi Feng, Junliang Xing, Shuicheng Yan
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 joint partition (pose partition) before inference (ablation) | PPN · MPII validation | total AP: 78.6 -> 79.0 | Improved |
| add joint partition (pose partition) before inference (ablation) | PPN · MPII validation | inference time (ms): 3.4 -> 1.9 | Improved |
| add local greedy inference (ablation) | PPN · MPII validation | total AP: 77.8 -> 79.0 | Improved |
| add single-person pose refinement (ablation) | PPN · MPII validation | total AP: 76.4 -> 79.0 | Improved |
| test at 384x384 instead of 256x256 (ablation) | PPN · MPII validation | total AP: 76.3 -> 79.0 | Improved |
| add refinement and multi-scale testing to vanilla PPN (ablation) | PPN · MPII validation | total AP: 74.8 -> 79.0 | Improved |
| use 8 stacked hourglass stages instead of 1 (ablation) | PPN · MPII validation | total AP: 64.0 -> 79.0 | Improved |
| remove partition: head and shoulder detection vs per-joint (ablation) | PPN · MPII validation | hip and knee AP: Hip 77.8 -> 78.8, Knee 72.7 -> 73.1, Ankle 64.7 -> 65.7 | Hurt |
| replace Newell and Deng associative embedding with PPN | bottom-up pose estimation · MPII multi-person (full test set) | total AP: 77.5 -> 80.4 | Improved |
| replace Cao et al. with PPN for speed | bottom-up pose estimation · MPII multi-person (full test set) | runtime (s): 1.24 -> 0.77 | Improved |
| replace Xia et al. with PPN | pose estimation · PASCAL-Person-Part (extended) | total AP: 39.2 -> 46.6 | Improved |
| replace Insafutdinov et al. with PPN | pose estimation · WAF | total AP: 82.0 -> 84.8 | Improved |
Closest papers by meaning
- MultiPoseNet: Fast Multi-Person Pose Estimation using Pose Residual Network (ECCV 2018 · similarity 0.90)
- DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation (CVPR 2016 · similarity 0.89)
- Graph and Temporal Convolutional Networks for 3D Multi-person Pose Estimation in Monocular Videos (AAAI 2021 · similarity 0.89)
- Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation (AAAI 2020 · similarity 0.88)
- PoseTrack: Joint Multi-Person Pose Estimation and Tracking (CVPR 2017 · similarity 0.88)
- DeeperCut: A Deeper, Stronger, and Faster Multi-person Pose Estimation Model (ECCV 2016 · similarity 0.88)
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.
