Cartoon Face Expression Simulation Based On Depth Image

The 3D facial expression animation is a research hotpoint and keypoint of computer graphics.In recent years,virtual character animations have appeared in the field of film and television production,video chat,online games and advertisement production more frenquently,which making virtual character animation occupy a wide space in production and life.Therefore,this paper proposes a simulation method based on depth image for cartoon facial expression animation.The 3D cartoon face model is used to simulate the facial expression of real person in the video.The real-time facial expression animation simulation was completed by this method.In this paper,the facial expression simulation system is mainly divided into three parts.Firstly,the depth image of human face is captured by the Kinect camera.The depth image is denoised and the foreground is extracted.What is more,the human face is identified and tracked in the acquisition process.The facial feature points' location coordinate in space are extracted;Secondly,the IFTModel class interface of Kinect is used to provide a generalized Candide-3 facial model for the user's face.The FaceGen is used to complete the construction of the 3D cartoon facial mesh routing and the OpenGL is used to complete texture mapping and rendering.According to MPEG-4standard,the cartoon facial feature points are signed by manual;Finally,the mapping is completed based on facial feature points to establish the correspondence between the two models,and then the alignment of the facial model is achieved through the deformation based on the Gaussian function interpolation,the nature of the Laplacian deformation algorithm can preserve the facial detail,which is employed to achieve facial expression simulation with adding Bézier curve constraints,.The experimental results show that the method in this paper can achieve real-time facial expression simulation and have a strong sense of reality.Compared with the facial expression simulation system based on 3D scanning data and motion capture technology,this paper has significantly improved the experimental cost and efficiency.Through comparison and analysis with other literatures,the advantages of this algorithm are summarized.