Reconstruct 3D human face model from single monocular image on mobile devices within milli-seconds. Seperate facial expressions from facial identities, and make them transferrable. Based on this technology, any transformations done on images can be seemlessly and stably transferred to videos.
Automatically separate people from background in videos on mobile devices in milli-seconds.
Combining with video segmentation, prettify and stylize foreground contents and background contents separately.