Natural skin color transformation based on color subspace learning

摘要

With the rapid spread of digital cameras, a growing interest has been seen in image processing for beautification. Pseudo make-up face can be generated by a variety of user's preferred make-up as the image processing method for beautification. But unnatural image be generated by this method because of different lighting conditions or skin colors between two images. Our proposed method can generate more natural images by transforming skin color which based on color subspace learning. We also developed a skin color transformation system, which can transform the effect of make-up rapidly and naturally.