Research map / 3D vision / Human pose estimation
PyMAF-X: Towards Well-Aligned Full-Body Model Regression From Monocular Images
Hongwen Zhang, Yating Tian, Yuxiang Zhang, Mengcheng Li, Liang An, Zhenan Sun, Yebin Liu
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 mesh-aligned feedback (PyMAF) over baseline, ResNet-50 | HMR-style baseline · 3DPW and Human3.6M | MPJPE: reduces MPJPE by 4.7 mm and 5.5 mm | Improved |
| add mesh-aligned feedback (PyMAF) over baseline, ResNet-50 | HMR-style baseline · COCO val keypoints | AP: 33.7 -> 42.2 | Improved |
| use mesh-aligned features instead of global features (ablation) | PyMAF · Human3.6M | MPJPE: 84.1 -> 79.6 | Improved |
| use mesh-aligned features instead of grid features (ablation) | PyMAF (pyramidal) · Human3.6M | MPJPE: 79.7 -> 76.8 | Improved |
| add feature pyramid to mesh-aligned feedback (ablation) | PyMAF · Human3.6M | MPJPE: 79.6 -> 76.8 | Improved |
| add dense correspondence auxiliary supervision (ablation) | PyMAF · Human3.6M | MPJPE: 76.8 -> 75.1 | Improved |
| use dense correspondence instead of part segmentation as supervision (ablation) | PyMAF · Human3.6M | MPJPE: 75.5 -> 75.1 | Improved |
| feed auxiliary predictions rather than features to regressor (ablation) | PyMAF · Human3.6M | MPJPE: 75.1 -> 77.8 | Hurt |
| add spatial alignment attention (ablation) | PyMAF · Human3.6M | MPJPE: 76.8 -> 73.6 | Improved |
| add grid features by fully-connected fusion instead of SAA (ablation) | PyMAF · Human3.6M | MPJPE: 76.8 -> 76.6 | No clear effect |
| add spatial alignment attention across refinement iterations (ablation) | PyMAF · Human3.6M | MPJPE: 75.1 -> 72.1 | Improved |
| use adaptive integration instead of learned integration (ablation) | PyMAF-X · EHF | hand MPJPE: 42.3 -> 31.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- OpenPose — OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields (TPAMI 2019) · better on 0 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
- PyMAF: 3D Human Pose and Shape Regression With Pyramidal Mesh Alignment Feedback Loop (ICCV 2021 · similarity 0.94)
- Recovering 3D Human Mesh From Monocular Images: A Survey (TPAMI 2023 · similarity 0.89)
- One-Stage 3D Whole-Body Mesh Recovery With Component Aware Transformer (CVPR 2023 · similarity 0.89)
- BLADE: Single-view Body Mesh Estimation through Accurate Depth Estimation (CVPR 2025 · similarity 0.87)
- AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimation (CVPR 2024 · similarity 0.86)
- ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization (ICML 2025 · similarity 0.86)
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.
