A new deep-learning algorithm learned how to replace a background in portrait photos.
The AugmentedPerception team released the technical papers on their project for SIGGRAPH 2021. The project is called Total Relighting: Learning to Relight Portraits for Background Replacement and is created for replacing backgrounds of photos using deep-learning algorithms.
"We propose a novel system for portrait relighting and background replacement, which maintains high-frequency boundary details and accurately synthesizes the subject's appearance as lit by novel illumination, thereby producing realistic composite images for any desired scene. Our technique includes foreground estimation via alpha matting, relighting, and compositing. We demonstrate that each of these stages can be tackled in a sequential pipeline without the use of priors (e.g. known background or known illumination) and with no specialized acquisition techniques, using only a single RGB portrait image and a novel, target HDR lighting environment as inputs," write the developers in the technical papers.