Single-image Full-body Human Relighting

Abstract
We present a single-image data-driven method to automatically relight images with full-body humans in them. Our framework is based on a realistic scene decomposition leveraging precomputed radiance transfer (PRT) and spherical harmonics (SH) lighting. In contrast to previous work, we lift the assumptions on Lambertian materials and explicitly model diffuse and specular reflectance in our data. Moreover, we introduce an additional light-dependent residual term that accounts for errors in the PRTbased image reconstruction. We propose a new deep learning architecture, tailored to the decomposition performed in PRT, that is trained using a combination of L1, logarithmic, and rendering losses. Our model outperforms the state of the art for full-body human relighting both with synthetic images and photographs.
Description

        
@inproceedings{
10.2312:sr.20211300
, booktitle = {
Eurographics Symposium on Rendering - DL-only Track
}, editor = {
Bousseau, Adrien and McGuire, Morgan
}, title = {{
Single-image Full-body Human Relighting
}}, author = {
Lagunas, Manuel
and
Sun, Xin
and
Yang, Jimei
and
Villegas, Ruben
and
Zhang, Jianming
and
Shu, Zhixin
and
Masia, Belen
and
Gutierrez, Diego
}, year = {
2021
}, publisher = {
The Eurographics Association
}, ISSN = {
1727-3463
}, ISBN = {
978-3-03868-157-1
}, DOI = {
10.2312/sr.20211300
} }
Citation