Research On Iris Rapid Location And Self-adaptive Recognition

Information security in modern society is the focus of attention,and accurate identification of individual identity is the key to ensure the safety.Therefore,individual unique biological information which is based on fingerprints,human face,iris,palm print and gait,is widely used in identity authentication technology,and the application of iris recognition technology has become a research focus for the high degree of uniqueness and anti-counterfeit of the iris image.As a perfect identity authentication technology not only need high accuracy,and its good broad-spectrum and real-time nature is also indispensable.In this paper,based on the analysis of the research status of iris recognition technology,we make some research on the rapid localization and self-adaptive recognition of irregular problems for concrete link.Firstly,according to the characteristics of iris which is round but non-circular,and the iris is often disturbed by eyelids and eyelashes,we present a fast iris localization method.The method uses the gray scale features of the eye image to locate the inner boundary accurately.Then,a new method of locating the outer boundary from coarse to fine is proposed.This method makes the inner boundary of the localization closer to the real boundary,and the outer boundary localization method greatly reduces the positioning time without affecting the accuracy of positioning with the circular difference algorithm.Moreover,the Canny operator is improved to reduce the number of false edges and achieve a rapid detection of eyelids edge.Secondly,the iris texture feature is extracted by using wavelet transform and gray level co-occurrence matrix.But the characteristics of the same eye's different iris images will be different,so we train and recognize the extracted features through the BP algorithm;For the slow convergence speed of BP neural network which leads to training time is longer and so on,we give two improved methods,the variable learning rate of neural network and accelerated BP algorithm of epsilon vector extrapolation method.Two improved methods not only reduce training time and improve efficiency,but also improve the recognition rate.Finally,based on a series of results obtained in this article,we conduct iris rapid positioning and self-adaptive identification system frame construction and design combined with MFC in Visual Studio 2010 development platform.