Face Cartoon Synthesis Based on the Active Appearance Model

In this paper, we propose a new method to synthesize a face cartoon from a face photo based on the Active Appearance Model (AAM). Compared with the current block-based methods, our algorithm matches the patches using the position of the facial feature points. The method consists of two steps, i.e., feature points searching and cartoon face synthesis. In the first step, feature points are constructed using the AAM which searches the outline of the facial features. In the second step, the hair is processed through image segmentation and image gradient, the contour is drawn directly according to the feature points obtained by the AAM, and the facial organs are synthesized by patches from the training database. This algorithm is trained and tested on the CUHK student database. Experimental results demonstrate that the proposed approach achieves a good performance compared with the state-of-art methods.