Too Long; Didn't Read
A new paper by Google presents a novel per-pixel lighting representation in a deep learning framework. This explicitly models the diffuse and the specular components of appearance, producing relit portraits with convincingly rendered effects like specular highlights. The most impressive thing about this paper is that these two tasks are made without any other information than two pictures: your portrait and the new background to create this new realistic image. This task is called “Portrait relighting and background replacement” and has two very complicated sub-tasks: Background replacement.
Share Your Thoughts