ETCH
Generalizing Body Fitting to Clothed Humans via Equivariant Tightness

International Conference on Computer Vision 2025 (ICCV 2025)
⭐ Highlight Paper
Boqian Li1, Haiwen Feng2,3*, Zeyu Cai1, Michael J. Black2, Yuliang Xiu1,2†
1Westlake University 2Max Planck Institute for Intelligent Systems 3UC Berkeley
*Project Lead †Corresponding Author
Code arXiv Twitter Talk Video Paper PDF ICCV Proceedings Poster
Project Teaser Image

Body Fitting on Clothed Humans. Given 3D clothed humans in any pose and clothing, ETCH accurately fits the body underneath. Our key novelty is modeling cloth-to-body SE(3)-equivariant tightness vectors for clothed humans, abbreviated as ETCH, which resembles "etching" from the outer clothing down to the inner body. The ground-truth body is shown in blue, our fitted body in green, and ground-truth markers as ●.

Video Demonstration of How ETCH Works

Clothed Human Fitting Results

(Scan, Fitted SMPL, 3D Tightness Hotmap)

Method

ETCH Pipeline

1. ETCH Pipeline: 1) Equivariant Tightness Vector Prediction, which takes the sampled points X as input, and estimates the tightness directions D via equivariant features fequiv , along with the tightness magnitudes B, labels L, and confidences C via invariant features finv . With these ingredients, in 2) Marker Aggregation and SMPL Optimization, the points move inward along the tightness vectors, forming body-shaped point clouds. These points are weighted and aggregated, based on their labels and confidences, to produce final markers for SMPL fitting.

Tightness Vector Illustration

2. Illustration of our Tightness-Vector and Marker-Confidence system.

Abstract

Fitting a body to a 3D clothed human point cloud is a common yet challenging task. Traditional optimization-based approaches use multi-stage pipelines that are sensitive to pose initialization, while recent learning-based methods often struggle with generalization across diverse poses and garment types. We propose Equivariant Tightness Fitting for Clothed Humans, or ETCH, a novel pipeline that estimates cloth-to-body surface mapping through locally approximate SE(3) equivariance, encoding tightness as displacement vectors from the cloth surface to the underlying body. Following this mapping, pose-invariant body features regress sparse body markers, simplifying clothed human fitting into an inner-body marker fitting task. Extensive experiments on CAPE and 4D-Dress show that ETCH significantly outperforms state-of-the-art methods -- both tightness-agnostic and tightness-aware -- in body fitting accuracy on loose clothing (16.7% ~ 69.5%) and shape accuracy (average 49.9%). Our equivariant tightness design can even reduce directional errors by (67.2% ~ 89.8%) in one-shot (or out-of-distribution) settings (≈1% data). Qualitative results demonstrate strong generalization of ETCH, regardless of challenging poses, unseen shapes, loose clothing, and non-rigid dynamics. We will release the code and models soon for research purposes at boqian-li.github.io/ETCH.

BibTeX

@inproceedings{li2025etch,
title = {{ETCH: Generalizing Body Fitting to Clothed Humans via Equivariant Tightness}},
author = {Li, Boqian and Feng, Haiwen and Cai, Zeyu and Black, Michael J. and Xiu, Yuliang},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2025},
pages = {8264--8274},
doi = {10.1109/ICCV51701.2025.00774}
}