An Improved Face Recognition Method Using Global Filled Function

3D data registration and classifier are two important components in face recognition system. Aiming at the current methods´ handicaps such as slow convergence and easiness of getting into local optimization, this paper presents a novel face recognition method using filled function, one of the effective deterministic methods. It gives a modified concept of filled function based on Ge, and proposes a more practicable one-parameter filled function. Then it works out an improved ICP 3D data registration algorithm and an improved BP neural network classifier, both combining filled function method. The filled function method can find a lower minimizer by leaving the minimizer previously found. By repeating these processes, a global minimizer is obtained. Experiments show that this face recognition method decreases the amount of calculation, improves the accuracy of recognition precision and has an actual recognition effect.