Research On The Technique Of 3D Face Modeling Based On Image

Building realistic 3D face models is a challenging task in the field of computer graphics. With the development of virtual characters of movies, remote conference, criminology, medical science, information help, human-machine interaction, amusement ,virtual reality , face recognition, understanding of expression ,etc, people are paying more attention to the face modeling area.Though it has already made the break-through on some technology in the reconstruction and duplication of 3D face at present, it's not very satisfying. Because it needs special-purpose appar- atus, the data amount is big and it is inconvenient to store and transmit through the network. How to establish the 3D face model with sense of reality without expensive apparatus and with less human-machine interaction is the keystone of our research, which aims at finding some simple and high-efficient methods to solve these problems.This paper paid attention to the research on the technique of 3D face modeling based on image. That is to say, rebuild 3D face with two orthogonal photos of person's front and profile faces. First, we should do pretreatment on the front and profile views of the user's face. Second, this paper presented a method of face detection based on the segmentation of skin color and then extracted the person face outline. After that we extracted the information of the feature points with method of wavelet analysis. Then, we obtained the specific face model with the specific people's face characteristic by modifying the common face model. Finally, a cylindrical texture map is built from the two image views to generate a realistic face image.