Frontal And Lateral Pictures Based Three-dimensional Facial Modeling

Generating realistic 3D human face models has always been a persistent challenger problem in computer graphics. The modeling and animation of realistic human face has almost 30 years as an identifiable area. With the development of virtue reality, movie making, visual telephone, etc. This research area is now receiving more attention than ever before. This thesis makes a deeply research on the area.This thesis gives the future application of 3D face models and the present research, carry through the compare and analysis with the 3D facial model technology in existence, base on the frontal and lateral human face pictures that input automatically, puts forward a synthetical linear spring model and DFFD based facial modeling method;This method defines the facial feature points on the research of human facial character;brings forward a new integrated method in geometrical structure reconstruction., At last, using mesh subdivision with the local human face, rendering eliminant with the special human face mesh that has been changed and subdivided, and getting the realistic eliminate image of special human face. The main research topics include:1) Defining facial feature points.The topic is to define the feature points of a real face. These facial feature points must show not only the whole outline of a face, but also the local character of the face components.2) Geometrical structure reconstruction. this method is a combination of DFFD and linear spring model method, using the two different methods to tackle non-character points movements in the region of sparseness or denseness with the character points, matching the normal face model with the given face model.3) Local area subdivision. through the subdivision of the facial denseness area constantly, we can keep the fine detail character of human face.This paper presented a feasible engineering facial modeling method. And this method has some practical significances.