Based On Improved Gabor Feature ExtractionLip-print Recognition

With the rapid development of information,network and digitization,biometrics has become more and more concerned with the advantages of congenital convenience and high recognition rate,such as fingerprint recognition,palmprint recognition,face recognition,iris recognition,ear Profile recognition and so on.As an emerging technology of biometrics,lip-print recognition technology has the advantages of high recognition precision,short recognition time,high user acceptance and low hardware cost.This paper designs and realizes the lip-print recognition system according to the characteristics of lip-print image.This paper based on thenon-contact way of mobile phone to collect lip-print images,According to the characteristics of lip-print color relative to other parts of the face skin red,the RGB lip image is transformed into YCbCr image and the Cr of the lip-print is highlighted and then used Sobel edge detection and morphological transform toobtain the collection of lip-print image positioning.For collection of lip-print image texture feature is not prominent question,this article puts forward the improved histogram equalization image enhancement algorithm,it is used butterworth high-pass filter to enhance high frequency characteristics of thelip-print image,then combined with the feature of histogram equalization to lip-print,The experimental results show that the improved image enhancement algorithm enhances the lip-print texture.Aiming at the problem that the feature dimension of Gabor filter is too large,an improved Gabor feature extraction algorithm is proposed,which combines Gabor filter and PCA principal component analysis method,and finally realizes the feature extraction and dimension reduction of lip-print image.The design and implementation of the lip-print recognition system are divided into registration phase and recognition stage.The registration phase includes image acquisition,lip-print localization,the image enhancement,the feature extraction of the lip-print and the establishment of the feature library.The recognition phase includes the lip-print localization of the lip-print image to be verified,the image enhancement,the feature extraction,the lip-print feature to be verified are matched by the Euclidean distance calculated from the feature in the lip-print feature library,and the minimum Euclidean distance is the identity of the verifier To identify the identity of the person.Finally,this article through the GUI interface to achieve a simple lip-print recognition system.