Facial Appearance Capture at Home with Patch-Level Reflectance Prior Journal: TOG Volume: 44 4 8 DOI: 10.1145/3730825 CCS: Computing methodologies Rendering
Yuxuan Han, Junfeng Lyu, Kuan Sheng, Minghao Que, Qixuan Zhang, Lan Xu, Feng Xu · 2025
Read Facial Appearance Capture at Home with Patch-Level Reflectance Prior Journal: TOG Volume: 44 4 8 DOI: 10.1145/3730825 CCS: Computing methodologies Rendering by Yuxuan Han, Junfeng Lyu, Kuan Sheng, Minghao Que, Qixuan Zhang, Lan Xu, Feng Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-03478.txt.
Description
Capturing a face well enough for film-quality digital doubles usually takes a studio rig of lights and cameras. Seven researchers close that gap at home in 2025 with a phone and its flashlight in a dim room. Their trick is a prior: a diffusion model trained on studio scans learns what real skin reflectance looks like, one small patch at a time. Patch training stretches a small dataset further and steadies the learning.
During capture, the system samples full-face reflectance maps patch by patch, steering each one to match the phone video. The seam-free maps come out close to studio quality in tests. The authors say everyday users could soon clone their own faces into the digital world with gear they already own.
Catalog record
| CREATOR | Yuxuan Han, Junfeng Lyu, Kuan Sheng, Minghao Que, Qixuan Zhang, Lan Xu, Feng Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-03478