CVPR2023

Controllable Light Diffusion for Portraits

David Futschik, Kelvin Ritland, James Vecore, Sean Fanello, Sergio Orts-Escolano, Brian Curless, Daniel Sýkora, Rohit Pandey

Abstract

Figure 1 . Our method enables control over the diffuseness of light in arbitrary portrait images. We first extract specular/shadow maps from the input image and then produce a fully diffuse image. Additionally, we show how to recover a uniformly lit image, i.e. tinted by the average light color, from which we can estimate the untinted albedo. The bottom row illustrates the application of editing the input photo by gradually increasing the amount of light diffusion.