An Automatic Virtual Makeup Scheme Based on Personal Color Analysis

Jisoo Park, Hyungjoon Kim, Seonmi Ji, Eenjun Hwang

Abstract: With people's timeless interest in beauty and the recent development of IT technologies, there has been a growing demand for a system that can suggest appropriate cosmetics or fashion styles to each individual. So far, professionals have made such suggestion subjectively or depending on personal experience, which produced inconsistent results. One possible method to do this objectively and systematically would be to utilize the personal color of an individual. Personal color, also known as skin tone color matching is the process of finding colors of clothing and makeup to match a person's skin complexion, eye color, and hair color. In this paper, we propose a scheme for automatically generating virtual makeup based on personal color. To do that, we first identify user's personal color from the image by analyzing the color of pupils, hair, and skin, and then select cosmetic method and colors. Lastly, by looking up the predefined makeup database for foundation, blush, lipstick, eyeline, and eyeshadow, we produce a virtual makeup on the user's face image.