Portrait beautification: A fast and robust approach

摘要

Portrait beautification technology has important significance regarding virtual background photography and augment reality said. In this work, a set of integrated portrait beautification approaches are raised. Aiming at fast portrait division, quantum evolution algorithm is introduced into the course of threshold searching. In order to merge images and virtual background, based on Human Vision System principles, a multi-layer portrait weighted method is proposed. To overcome the effects of illumination on face skin color, one kind of fast face color correction technology based on color temperature estimation is presented. Detail experimental results and comparisons are provided to show our approaches are effective.