Research map / 3D vision / Human pose estimation
VirtualPose: Learning Generalizable 3D Human Pose Models from Virtual Data
Jiajun Su, Chunyu Wang, Xiaoxuan Ma, Wenjun Zeng, Yizhou Wang
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 3D root estimator (REN 2D+3D) on top of 2D depth-map root estimation (ablation) | VirtualPose · MuPoTS-3D | MRPE_z (mm): 245.4 -> 165.0 | Improved |
| add 3D root estimator (REN 2D+3D) on top of 2D depth-map root estimation (ablation) | VirtualPose · CMU Panoptic | MRPE_z (mm): 104.1 -> 93.6 | Improved |
| add Pose Estimation Network to refine root joint (ablation) | VirtualPose · MuPoTS-3D | MRPE_z (mm): 165.0 -> 159.8 | Improved |
| add Pose Estimation Network to refine root joint (ablation) | VirtualPose · CMU Panoptic | MRPE_z (mm): 93.6 -> 86.0 | Improved |
| add Pose Estimation Network to refine root joint (ablation) | VirtualPose · CMU Panoptic | MRPE (mm): 115.7 -> 97.0 | Improved |
| add 3D root estimator (REN 2D+3D) (ablation) | VirtualPose · CMU Panoptic | MRPE (mm): 113.9 -> 115.7 | Hurt |
| add depth-based feature projection instead of projecting heatmaps to all voxels (ablation) | VirtualPose · MuPoTS-3D | MRPE (mm): 253.2 -> 204.0 | Improved |
| add depth-based feature projection instead of projecting heatmaps to all voxels (ablation) | VirtualPose · MuPoTS-3D | MRPE_z (mm): 210.5 -> 159.8 | Improved |
| train on virtual data from uniformly random cameras instead of a single camera (ablation) | VirtualPose · CMU Panoptic cross-view test camera 14 | MRPE_z (mm): 454.6 -> 273.8 | Improved |
| generate training data from the known test camera 19 viewpoint (ablation) | VirtualPose · CMU Panoptic camera 19 | MRPE_z (mm): 155.2 -> 62.3 | Improved |
| train with poses taken from MuPoTS-3D test set (ablation) | VirtualPose · MuPoTS-3D | PCK_abs: 44.0 -> 50.4 | Improved |
| add augmented backgrounds to training data (ablation) | BDR baseline · MuCo-3DHP augmented-background test | MRPE_z (mm): 267.8 -> 148.9 | Improved |
Closest papers by meaning
- AdaptPose: Cross-Dataset Adaptation for 3D Human Pose Estimation by Learnable Motion Generation (CVPR 2022 · similarity 0.90)
- Benchmarking and Analyzing 3D Human Pose and Shape Estimation Beyond Algorithms (NeurIPS 2022 · similarity 0.89)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.88)
- A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation (CVPR 2024 · similarity 0.88)
- Can Generative Video Models Help Pose Estimation? (CVPR 2025 · similarity 0.88)
- Multiple View Geometry Transformers for 3D Human Pose Estimation (CVPR 2024 · 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.
