Surface Map And Registration Method Based On Conformal Geometric

The matching of three dimension surface and recognition technology is a hot research topic in computer graphics field. With the advent of the three dimension scanner, surface and recognition technology have been carried out extensive research in many disciplines and fields, such as computer vision, texture mapping, face matching and registration, shape analysis, image processing and pattern recognition, airport and station inspection, and so on. However, most of the researches of face recognition are based on the two dimensional plane images in the actual application. It is less for the researches of three dimensions surface. This paper introduced conformal geometry method to study surface matching and registration, it is very convenient to convert three dimension facial surface into two-dimensional plane, At last, a novel algorithm of three dimension facial expression recognition is put forward.Firstly, this paper introduces common surface mapping of conformal geometry, then introduces discrete Ricci flow as an efficient and powerful mathematics tools and the surface conformal parameterization algorithm is designed and implemented.Making use of the algorithm of the first step, the face surface is mapped to the plane area, then make a diffeomorphism between planar regions by using harmonic maps, so as to realize the matching and registration of surfaces.The three dimension facial expression recognition algorithm based on the Ricci curvature flow is established by making use of the algorithm of the first and two steps. The face surface is conformal mapped to standard plane,so, conformal modules of facial surface is given with circle center and radius of the inner circle, The rate of change for each point area (conformal factor) gives the Riemann metric, and by compared the coordinate surfaces of the Teichmuller space and average conformal modules, so as to realize the facial expression classification algorithm of three dimension facial surface.The algorithm of the design and implementation is applied to three dimension face scan of real data, then, the three dimension facial surface is processed by conformal parameterization, registration and it is compared with average conformal mode, so as to realize the classification of facial expression.