Applied Research On Intelligent Recognition Based On Fuzzy Technology

Face recognition involves many fields including intelligent recognition, image processing, physiology, mathematics and so on, it has important theoretical value and great practical value, can be widely used for security, authentication and other fields, and is hot topic in the field of biometrics. Face recognition is one of the most important for identity authentication in the technology of biometrics, and be attention by more and more people for its directness, friendshop and convenience compared with other identification. Due to various factors, face recognition has always been considered a subject of great challenge(1)The number of fuzzy rules in traditional fuzzy neural network is very larger when dealing with high dimensional data, so the fuzzy rules must be streamlined for face recognition using fuzzy neural network method. By extracting the mean and variance of each dimension in each class of face image, and using the product of the membership that each dimension of face image belogns to each class of face images for the degree of application of fuzzy rules, the number of fuzzy rules can be streamlined to equal the number of the classes of samples, thus the fuzzy neural network can be applied to face recognition that need deal with high dimensional data.(2)This paper has implemented face recognition base on fuzzy rules streamlined fuzzy neural network,and has made a lot of expriments on ORL face database and ORL face database added small amount of Gaussian nosie.Experimental results show that face recognition base on fuzzy rules streamlined fuzzy neural network has good recognition rate and anti-noise ability.(3)On the basis of fuzzy rules streamlined, this paper has proposed the method of face recognition based on interval type-2 fuzzy neural network by combining type-2 fuzzy sets with neural network, has implemented face recognition base on internal type-2 fuzzy neural network, and has made a lot of expriments on ORL face database and ORL face database added small amount of Gaussian nosie. Experimental results show that face recognition based on interval type-2 fuzzy neural network has a good effect in recognition rate and anti-noise ability.