Method of automatic face recognition based on fuzzy multiple feature combination

In general, automatic face recognition technique includes two important steps: detection/location, recognition. In this paper, a method of automatic face recognition based on fuzzy multiple feature combination was proposed. Face detection and location used eigen space and gray image understand processing. The segmented face image was recognized by the methods: Eigenfaces, EigenUpper, EigenTzone and two-order Eigenfaces firstly. The fuzzy integration function was adopted to combine the fuzzy results, which were fuzzy fused from the elementary recognition results, in order to obtain new distance function and the final recognition results were given. The experiments on the Yale and ORL face database show that, the correct recognition rates of the method are all greater than 95% and the fuzzy multiple feature combination method is superior to that of classical eigenfaces method.