Research map / 3D vision / Face and shape
ETCH: Generalizing Body Fitting to Clothed Humans via Equivariant Tightness
Boqian Li, Haiwen Feng, Zeyu Cai, Michael J. Black, Yuliang Xiu
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 equivariant features instead of XYZ positions for direction (ablation) | ETCH · 4D-Dress | V2V: 2.033 -> 1.939 | Improved |
| Use equivariant features instead of XYZ plus invariant features (ablation) | ETCH · 4D-Dress | V2V: 2.307 -> 1.939 | Improved |
| Use equivariant features instead of invariant features only (ablation) | ETCH · 4D-Dress | V2V: 2.842 -> 1.939 | Improved |
| Use equivariant features instead of XYZ on CAPE (ablation) | ETCH · CAPE | V2V: 1.661 -> 1.647 | No clear effect |
| Use equivariant features in one-shot training (0.66% of 4D-Dress) (ablation) | ETCH · 4D-Dress one-shot | median angular error: 0.988 -> 0.315 | Improved |
| Use equivariant features in one-shot training (2.1% of CAPE) (ablation) | ETCH · CAPE one-shot | median angular error: 0.527 -> 0.0535 | Improved |
| Use sparse markers instead of dense correspondence (ablation) | ETCH · CAPE | V2V: 1.909 -> 1.647 | Improved |
| Use sparse markers instead of dense correspondence (ablation) | ETCH · 4D-Dress | V2V: 2.285 -> 1.939 | Improved |
| Add tightness direction vector to scalar under dense correspondence (ablation) | ETCH · CAPE | V2V: 1.777 -> 1.909 | Hurt |
| Add tightness direction vector to scalar under dense correspondence (ablation) | ETCH · 4D-Dress | V2V: 2.410 -> 2.285 | Improved |
| Add chamfer-based post-refinement (ablation) | ETCH · CAPE | V2V: 1.647 -> 1.017 | Improved |
| Add chamfer-based post-refinement (ablation) | ETCH · 4D-Dress | V2V: 1.939 -> 3.474 | Hurt |
Closest papers by meaning
- ETCH-X: Robustify Expressive Body Fitting to Clothed Humans with Composable Synthetic Data (ECCV 2026 · similarity 0.95)
- SCALE: Modeling Clothed Humans with a Surface Codec of Articulated Local Elements (CVPR 2021 · similarity 0.89)
- Cloth2Body: Generating 3D Human Body Mesh from 2D Clothing (ICCV 2023 · similarity 0.88)
- OmniFit: Multi-modal 3D Body Fitting via Scale-agnostic Dense Landmark Prediction (ECCV 2026 · similarity 0.88)
- Garment Recovery with Shape and Deformation Priors (CVPR 2024 · similarity 0.87)
- VS: Reconstructing Clothed 3D Human from Single Image via Vertex Shift (CVPR 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.
