Research map / 3D vision / Face and shape
Tex2Shape: Detailed Full Human Body Geometry From a Single Image
Thiemo Alldieck, Gerard Pons-Moll, Christian Theobalt, Marcus Magnor
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 MS-DSSIM loss to L1 and GAN losses (ablation) | Tex2Shape · Own synthetic test set | clothing reconstruction quality (visual) | Improved |
| use DensePose UV mapping instead of ground truth UV mapping (ablation) | Tex2Shape · Own synthetic test set | reconstruction quality (visual) | No clear effect |
| use HMR-based 3D pose texture reprojection instead of DensePose mapping (ablation) | Tex2Shape · Own synthetic test set | reconstruction quality (visual) | Hurt |
| add illumination augmentation during training (vs constant ambient light) (ablation) | Tex2Shape · Own synthetic test set | detail consistency (visual) | Improved |
| use 512x512 UV maps instead of 256x256 (ablation) | Tex2Shape · Own synthetic test set | detail quality (visual) | Improved |
| increase partial texture occupancy (visible pixels) | Tex2Shape (GT UV) · 55 test subjects, 3 poses | mean per-pixel displacement error | Improved |
| rotate subject around Y-axis beyond 30 degrees | Tex2Shape · A-pose test subjects | mean per-pixel displacement error: error identical from 0 to 30 degrees, increases linearly after 30 degrees | Hurt |
| vary focal length away from training value | Tex2Shape · A-pose test set | SMPL vertex-to-vertex error | No clear effect |
Closest papers by meaning
- 3DPeople: Modeling the Geometry of Dressed Humans (ICCV 2019 · similarity 0.89)
- Accurate 3D Body Shape Regression Using Metric and Semantic Attributes (CVPR 2022 · similarity 0.88)
- Detailed Human Shape Estimation From a Single Image by Hierarchical Mesh Deformation (CVPR 2019 · similarity 0.88)
- BodyMap: Learning Full-Body Dense Correspondence Map (CVPR 2022 · similarity 0.87)
- BodyNet: Volumetric Inference of 3D Human Body Shapes (ECCV 2018 · similarity 0.87)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · 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.
