一种人脸卡通画自动生成算法

he cartoon face parameter model defines the layout of face cartoon and its rendering in multiple layers through the reference shape and the painting pattern. A group of typical face cartoon sample drawn by artist can be parameterized into the painting style and the painting entity set using the model, and collected to constitute the cartoon sample space. In this paper, an automatic face cartoon synthesis algorithm based on sample parameter learning is proposed. Using deformation methods, the algorithm adjusts rendering shapes to geometric characteristics extracted from input facial image; with rendering elements, rendering rules of painting entities learned from sample space and its painting style, the algorithm automatically constructs a cartoon face parameter model for input face and generates the personalized face cartoon. Experiment and online application show that generated face cartoon preserves original facial characteristics, and its style is consistent with that of learning samples