We present a novel, zero-shot pipeline for creating hyperrealistic, identity-preserving 3D avatars from a few unstructured phone images. Existing methods face several challenges: single-view approaches suffer from geometric inconsistencies and hallucinations, degrading identity preservation, while models trained on synthetic data fail to capture high-frequency details like skin wrinkles and fine hair, limiting realism. Our method introduces two key contributions: (1) a generative canonicalization module that processes multiple unstructured views into a standardized, consistent representation, and (2) a transformer-based model trained on a new, large-scale dataset of high-fidelity Gaussian splatting avatars derived from dome captures of real people. This "Capture, Canonicalize, Splat" pipeline produces static quarter-body avatars with compelling realism and robust identity preservation from unstructured photos.
@article{arxiv.2510.14081,
title = {Capture, Canonicalize, Splat: Zero-Shot 3D Gaussian Avatars from Unstructured Phone Images},
author = {Emanuel Garbin and Guy Adam and Oded Krams and Zohar Barzelay and Eran Guendelman and Michael Schwarz and Matteo Presutto and Moran Vatelmacher and Yigal Shenkman and Eli Peker and Itai Druker and Uri Patish and Yoav Blum and Max Bluvstein and Junxuan Li and Rawal Khirodkar and Shunsuke Saito},
journal= {arXiv preprint arXiv:2510.14081},
year = {2025}
}
Comments
This work received the Best Paper Honorable Mention at the AMFG Workshop, ICCV 2025