3d Facial Animation Generation Method Design And Implementation

Realistic human facial animation is an active research field in computer graphics, which is widely used in movie making, pc games, video conference, man-machine interface, assistant teaching, remote commucation, medical research and so on.In this thesis, we propose a 3D facial expression animation framework based on facial feature points driving algorithm, which aims to achieve realistic feeling and real time performance.The main contribution of this paper includes:(1)analyse techniques of facial animation model. After analyzing three kinds of crucial technical in facial animation: geometric description, expression definition and mesh deformation technique, we chose high resolution face model and Laplacian mesh deformation technique in our system for realistic purpose.(2)Definition of isoline-based 3D facial motion model.We construct 3D facial motion model based on MPEG-4 standard and isoline-based mesh deformation technique.This new model has three advantage: Firstly, automatic search ROI region.The model automatic calculates the harmonic field of feature point to represent the ROI region, which is more smart than other model that with much manual work.Secondly,less time cost. The model uses the isoline set represent the reduced model of original mesh, so the original time consume problem of Laplacian mesh is removed. At last,more realistic expression result. It incorporates isoLine constraints into Laplacian reconstruction, which leads more reasonable and realistic result.(3)Implementation of facial animation system. We develop a 3D face animation system with isoline-based facial motion model. In our system, user can construct a facial motion model by selection a few feature points in high resolution 3D face model and use FAP stream to drive facial motion model. Experiments prove the proposed framework able to run in real time while satisfying the realistic requirement.We use MPEG-4 to define the moving of facial expression feature points to drive the 3D facial motion model that we defined in this paper. The experiment results demonstrate that the 3D facial motion model we defined can be used aminate the facial expression.