Face Recognition Based On 3D Modeling System

3D facial recognition is the current image processing of a relatively new research direction. In human computer authentication system of automatic fingerprint, genetic, iris and other methods to identify means of contact, need human samples, is invasive of identification. The recognition of non-contact characteristics of the computer field so as a research hot. At home and abroad has aroused high attention, has also made a series of results. 3D face recognition is a relatively new face recognition method is primarily used to solve 2D face recognition as difficult to solve brought people face because of posture, facial expressions, illumination and other factors which affect the efficiency of identifying an effective way. At home and abroad in research in this area was restricted to improve the recognition rate theory, how to find a basis for improve- ng the recognition rate while increasing the efficiency of the method of identification, so that 3D facial recognition as soon as possible into the practical state of the 3D faci- al recognition as the focus of and one of the difficulties.This paper deals with three-dimensional face modeling and 3D face recognition methods were compared and research, based on the model Candide-3 model three dimensional deformation of the method of face modeling, this method is relatively simple principle, easy to implement, while providing a very good three-dimensional face modeling process to improve the realistic level. Using global and local methods of combining three-dimensional face recognition, face recognition for the settlement process in the three-dimensional face model for three-dimensional data volume, com- putational complexity, real-time effects have a very good help. For the three dimensi- onal face recognition, this paper, the first cluster and then split three- dimensional face model to find features, and finally face to be identified according to the attitude adju- stment of the state of three-dimensional face model of a particular state of the perfor- mance of three-dimensional human face model to solve global structure The facial features of the mass center consisting of four point features, and performance of the surface of mouth, eyes, nose shape of the four key sub-regional characteristics of the eight-dimensional curvature, to match them to achieve their treatment of two dimensi- onal face image identification recognition.This three-dimensional facial reconstruction and recognition of the realization of Visual C6.0 and MATALB as the development tool, uses OpenGL technology to achi- eve the three-dimensional reconstruction, and achieved good results. Experiments show that, by presenting some of the improved method is feasible, experimental resul- ts can create realistic three-dimensional human face model, and for three-dimensional face recognition efficiency upgrade certain.