Research map / 3D vision / Human pose estimation
Keep It SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image
Federica Bogo, Angjoo Kanazawa, Christoph Lassner, Peter Gehler, Javier Romero, Michael J. Black
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 |
|---|---|---|---|
| replace the single-Gaussian pose prior with the mixture-of-Gaussians prior (ablation) | SMPLify · HumanEva-I (mean) | 3D joint error (mm): 103.7 -> 79.9 | Improved |
| add the interpenetration term to the single-Gaussian-prior model (ablation) | SMPLify · HumanEva-I (mean) | 3D joint error (mm): 104.1 -> 103.7 | No clear effect |
| use SMPLify instead of Zhou et al. on HumanEva-I | SMPLify · HumanEva-I (mean) | 3D joint error (mm): 110.0 -> 79.9 | Improved |
| use SMPLify instead of Zhou et al. on HumanEva-I | SMPLify · HumanEva-I (median) | 3D joint error (mm): 98.9 -> 61.9 | Improved |
| use SMPLify instead of Zhou et al. on Human3.6M | SMPLify · Human3.6M (mean) | 3D joint error (mm): 106.7 -> 82.3 | Improved |
| use SMPLify instead of Zhou et al. on Human3.6M | SMPLify · Human3.6M (median) | 3D joint error (mm): 90.0 -> 69.3 | Improved |
| use SMPLify instead of Zhou et al. on the SitDown action | SMPLify · Human3.6M (SitDown) | 3D joint error (mm): 137.5 -> 137.3 | No clear effect |
| add interpenetration term for complex poses (ablation) | SMPLify · LSP (qualitative) | qualitative pose plausibility | Improved |
| fit with more 2D joints (4 torso, 12 torso and limbs, 23 SMPL joints) (ablation) | SMPLify · Synthetic data (known pose) | shape error | Improved |
| fit shape from noisy 2D joints (1 to 5 px noise) instead of guessing the mean shape | SMPLify · Synthetic data | vertex-to-vertex shape error | Improved |
| use the gender-neutral model instead of gender-specific models (ablation) | SMPLify · LSP (qualitative) | visual pose quality | No clear effect |
Closest papers by meaning
- Probabilistic 3D Human Shape and Pose Estimation From Multiple Unconstrained Images in the Wild (CVPR 2021 · similarity 0.92)
- Skeleton-Aware 3D Human Shape Reconstruction From Point Clouds (ICCV 2019 · similarity 0.92)
- STAR: Sparse Trained Articulated Human Body Regressor (ECCV 2020 · similarity 0.92)
- Accurate 3D Body Shape Regression Using Metric and Semantic Attributes (CVPR 2022 · similarity 0.91)
- Expressive Body Capture: 3D Hands, Face, and Body From a Single Image (CVPR 2019 · similarity 0.91)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.89)
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.
