Makeup based on segmentation and local transfer

Abstract

In this paper, we introduce an automatic beauty system based on semantic segmentation and local transfer. The user provides a source image and a target image as a reference. Our system create a new makeup upon the face of source image with target image as reference. More precisely, we transfer each makeup areas(lip gloss, eye shadow, eyebrow)to corresponding parts of the face. One major advantages of our approach is we create this new makeup for source image locally, this not only caters to the requirement of users but also renders makeup upon face of source image more naturally. The effectiveness of the proposed are demonstrated in the experiments.