Iris Recognition Method Based On Data Pre Analysis

Now the application of iris recognition technology is more and more widely,such as finance,access control have been used;but with the application of iris recognition technology deeply,is also accompanied by many problems affecting the iris recognition rate;the solution for this problem is to improve the quality of iris image acquisition,using or developing better image processing algorithms.However,this kind of method in manpower,material and financial resources consumption;and this paper argues that the recognition rate is not ideal because of an iris image itself can not meet the recognition conditions,resulting in matching algorithm can not play the recognition performance.Therefore,in this paper,the pre classifier is used to select the image which is not suitable for matching.The main contents of this paper are: firstly,the iris image preprocessing,including iris and spot positioning and cutting,image mosaic,and then using 2D Gabor wavelet feature extraction,using the Euclidean distance matching;then the image matching was analyzed.The correct matching of iris image and matching error as two kinds of image samples,the two kinds of iris images as training samples by using gray co-occurrence matrix to extract the iris image,as image feature data,the characteristic data as the training set of support vector machine classifier is used to train pre classifier,and then the other samples tested;finally the support vector the trained classifier in machine recognition,image classification and recognition algorithm using the same pre made The results are compared with those of two times before and after recognition,and the feasibility of the method is proved by the two recognition matching rate.This paper mainly carries on the pre classification of iris image,through the training of pre classifier,support vector machine classification using characteristic two,selected does not apply in the back-end of iris image matching algorithm,so as to improve the recognition rate in the premise of the same recognition algorithm under.