Structure-Preserved Face Cartoonization

Abstract

Face cartoon synthesis has been proved to show wide range of uses in lots of fields, for example, instant message communication, suspects identity and online entertainment. In this paper, we propose a new dense descriptor based model to synthesize a face cartoon from a face photo which displays a great outcome. We generate two kinds of stylized face cartoons, one is called cartoon portraits and the other is called cartoon sketch. By integrating the two kinds of cartoons using guided filter, our results preserve the detail information of the input photo and generate good cartoon artistic style.