Research map / 3D vision / Human pose estimation
Not All Parts Are Created Equal: 3D Pose Estimation by Modeling Bi-Directional Dependencies of Body Parts
Jue Wang, Shaoli Huang, Xinchao Wang, Dacheng Tao
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 progressive bi-directional dependency modeling among body parts (ablation) | Baseline network · Human3.6M Protocol #1 | MPJPE (mm) average: 58.3 -> 56.9 | Improved |
| add predicted pose attributes as input (ablation) | Progressive network · Human3.6M Protocol #1 | MPJPE (mm) average: 56.9 -> 52.6 | Improved |
| add pose attributes and progressive modeling (ablation) | Baseline network · Human3.6M Protocol #1 | MPJPE (mm) wrist: 98.5 -> 87.6 | Improved |
| train attribute predictor on mixture of MPII and Human3.6M instead of Human3.6M only (ablation) | attribute predictor · Human3.6M attribute accuracy | accuracy: 78.6 -> 82.7 | Improved |
| add domain adaptation to mixed training (ablation) | attribute predictor · Human3.6M attribute accuracy | accuracy: 82.7 -> 84.0 | Improved |
| train on mixture with MPII instead of Human3.6M only (ablation) | attribute predictor · MPI-INF-3DHP attribute accuracy (unseen) | accuracy: 51.1 -> 67.7 | Improved |
| add domain adaptation to mixed training (ablation) | attribute predictor · MPI-INF-3DHP attribute accuracy (unseen) | accuracy: 67.7 -> 70.1 | Improved |
| use Ours instead of Dabral et al. | 3D pose estimator · Human3.6M Protocol #1 | MPJPE (mm) average: 55.5 -> 52.6 | Improved |
| use Ours instead of Pavlakos et al. | 3D pose estimator · Human3.6M Protocol #2 | MPJPE (mm) average: 41.8 -> 40.7 | Improved |
| use Ours instead of best prior method | 3D pose estimator · MPI-INF-3DHP | 3DPCK: 71.9 vs 71.9 | No clear effect |
| use Ours instead of best prior method | 3D pose estimator · MPI-INF-3DHP | AUC: 35.3 -> 35.8 | Improved |
| use multi-task network instead of pretrained hourglass | 2D pose detection · MPII validation | PCKh@0.5 average: 87.6 -> 88.5 | Improved |
Closest papers by meaning
- Learning Basis Representation to Refine 3D Human Pose Estimations (AAAI 2019 · similarity 0.89)
- Kinematic-Structure-Preserved Representation for Unsupervised 3D Human Pose Estimation (AAAI 2020 · similarity 0.89)
- Keep It SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image (ECCV 2016 · similarity 0.89)
- 3D Human Pose Estimation From a Single Image via Distance Matrix Regression (CVPR 2017 · similarity 0.88)
- A Sequential Approach to 3D Human Pose Estimation: Separation of Localization and Identification of Body Joints (ECCV 2016 · similarity 0.88)
- HiPoser: 3D Human Pose Estimation with Hierarchical Shared Learning at Parts-Level Using Inertial Measurement Units (AAAI 2025 · 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.
