Research map / 3D vision / Human pose estimation
Joint Optimization for Multi-Person Shape Models from Markerless 3D-Scans
Samuel Zeitvogel, Johannes Dornheim, Astrid Laubenheimer
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 |
|---|---|---|---|
| train a unisex model instead of a female-only model (ablation) | joint optimization shape model · FAUST intra-subject challenge | average error (cm): 2.353 -> 2.304 | Improved |
| train a unisex model instead of a female-only model (ablation) | joint optimization shape model · FAUST inter-subject challenge | average error (cm): 3.234 -> 3.525 | Hurt |
| train a male-only model instead of female-only (ablation) | joint optimization shape model · FAUST inter-subject challenge | average error (cm): 3.234 -> 3.519 | Hurt |
| combine female and male models instead of female only (ablation) | joint optimization shape model · FAUST intra-subject challenge | average error (cm): 2.353 -> 2.301 | Improved |
| combine female and male models instead of female only (ablation) | joint optimization shape model · FAUST inter-subject challenge | average error (cm): 3.234 -> 3.422 | Hurt |
| use ours (female) instead of LBS-AE | FAUST inter-subject · FAUST inter-subject challenge | average error (cm): 4.079 -> 3.234 | Improved |
| use ours (female) instead of unsupervised 3D-CODED | FAUST inter-subject · FAUST inter-subject challenge | average error (cm): 4.835 -> 3.234 | Improved |
| use ours (female) instead of supervised 3D-CODED | FAUST inter-subject · FAUST inter-subject challenge | average error (cm): 2.878 -> 3.234 | Hurt |
| use Deprelle et al. instead of ours (female) | FAUST intra-subject · FAUST intra-subject challenge | average error (cm): 2.353 -> 1.626 | Hurt |
| add shape blend-shape principal components from 0 to 10 (ablation) | trained female model · training scans | reconstruction error (mm): 3.7 mm with all 10 components | Improved |
Closest papers by meaning
- Learning to Estimate 3D Human Pose and Shape From a Single Color Image (CVPR 2018 · similarity 0.90)
- Probabilistic 3D Human Shape and Pose Estimation From Multiple Unconstrained Images in the Wild (CVPR 2021 · similarity 0.90)
- Hierarchical Kinematic Probability Distributions for 3D Human Shape and Pose Estimation From Images in the Wild (ICCV 2021 · similarity 0.89)
- Convolutional Mesh Regression for Single-Image Human Shape Reconstruction (CVPR 2019 · similarity 0.88)
- Human Body Model Fitting by Learned Gradient Descent (ECCV 2020 · similarity 0.88)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.87)
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.
