Research map / 3D vision / Face and shape
BaRe-ESA: A Riemannian Framework for Unregistered Human Body Shapes
Emmanuel Hartman, Emery Pierson, Martin Bauer, Nicolas Charon, Mohamed Daoudi
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 |
|---|---|---|---|
| Use Riemannian Sobolev metric on latent space instead of a linear (Euclidean) latent model (ablation) | BaRe-ESA · D-FAUST extrapolation | Chamfer distance: 0.44 -> 0.10 | Improved |
| Use Riemannian Sobolev metric on latent space instead of a linear (Euclidean) latent model (ablation) | BaRe-ESA · D-FAUST interpolation | Hausdorff distance: 0.18 -> 0.07 | Improved |
| Use BaRe-ESA instead of autoencoder baselines for registration of unseen scans | FAUST registration · FAUST (trained on D-FAUST) | Varifold distance: 0.021 -> 0.014 (vs ARAPReg) | Improved |
| Use BaRe-ESA instead of 3D-Coded on Hausdorff distance | FAUST registration · FAUST (trained on D-FAUST) | Hausdorff distance: 0.07 -> 0.08 | Hurt |
| Replace LIMP/ARAPReg with BaRe-ESA for extrapolation | D-FAUST extrapolation · D-FAUST extrapolation | Varifold distance: 0.082 -> 0.023 (vs ARAPReg) | Improved |
| Restrict optimization to a low-dimensional learned deformation basis instead of unrestricted H2 matching (ablation) | BaRe-ESA · qualitative arm movement comparison | naturalness of deformation | Improved |
| Use BaRe-ESA instead of deep autoencoders in terms of training time | BaRe-ESA · Training cost | training time: 2w -> <1h (vs ARAPReg) | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- 3D-Coded — 3D-CODED: 3D Correspondences by Deep Deformation (ECCV 2018) · better on 3 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Basis Restricted Elastic Shape Analysis on the Space of Unregistered Surfaces (IJCV 2024 · similarity 0.91)
- ATLAS: Decoupling Skeletal and Shape Parameters for Expressive Parametric Human Modeling (ICCV 2025 · similarity 0.86)
- BLSM: A Bone-Level Skinned Model of the Human Mesh (ECCV 2020 · similarity 0.85)
- COSMU: Complete 3D human shape from monocular unconstrained images (ECCV 2024 · similarity 0.85)
- MeasureXpert: Automatic Anthropometric Measurement Extraction from Two Unregistered, Partial, Posed, and Dressed Body Scans (ICCV 2025 · similarity 0.85)
- Numerical Inversion of SRNF Maps for Elastic Shape Analysis of Genus-Zero Surfaces (TPAMI 2017 · similarity 0.85)
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.
