3d Face Modeling Techniques Discussed

With the development of society and the improvement of science and technology, it is urgently needed for convenient and reliable automatic-status-discrimination. In this environment, the biometrics technique especially face recognition technique has become the highlight of machine intelligence research. The study of face recognition can promote the theory of image processing and pattern recognition in the status-discrimination and the searches based on contents. On the other hand, because of the particularity of the face pattern, the research of face recognition has the positive effect on the research about other sciences such as cognitive science, physiology and psychology.First of all, this dissertation summarizes the characteristic of biology recognition and its practical problems when implementation, and then the contents, the area and the history of face recognition are also elaborated. In the base of discussing the difficulties and challenges of 2D face recognition and 3D face recognition, this dissertation makes some work in 3D facial modeling, the first step of 3D face recognition. The main content of this dissertation is summarized below:(1) The meaning of biologic recognition, the choosing of the biologic characteristic and the application problems are discussed deeply. Around the face recognition, the progress of the face recognition research around the world is followed, and the application progress of the face recognition is investigated. In this dissertation, the main methods of traditional 2D face recognition and 3D face recognition are discussed, and the difficulties and challenges are also elaborated separately.(2) The problem of 3D facial modeling in 3D face recognition is discussed. It is the first step in 3D face recognition, in which the precision and integrality of 3D facial data decides the following work. The common methods of using 2D images to compose the 3D facial data are discussed. Because the human face has an unregular 3D structure, any method that uses two images to compose 3D facial data is doomed to have the shortcoming of inaccurate data, and reduce the precision of 3D face recognition. In the base of discussing the common two methods of composing 3D face, this dissertation produces two methods respectively that uses more images in order to improve the precision and integrality of the 3D facial data.