Algorithm Reserch Of Iris Recognition Based On 2D LogGabor Filter

With the rapid development of information technology, the research on information security has become one of important topics. The technology of biometrics identification has been receiving extensive attention by both scientists and researchers. Iris recognition, a kind of biological recognition, springs up in the last few years. As an important identification method, iris has characteristics of uniqueness, stability, collectible, noninvasiveness etc. In contrast with the other biometrics identification, it has the following advantages: more reliability, more stability, more security etc. Therefore, it has been agreed that the iris recognition technology is most promising one of the biometric recognition technology in the future, and it has wild respect and huge marketing.Iris identification refers to many fields like pattern identification, digital signal processing, image processing etc. The iris recognition system includes four parts: image acquisition, iris image preprocessing, feature extraction and matching. Iris image for this paper is provided by Institute of Automation Chinese Academy of Science (CASIA-Ver1).Iris feature extraction is based on texture analysis using 2D LogGabor filtering, after the research of abundant related literature and plentiful experiments. Its primary contributions in this paper are as follows:1)An improved algorithm is presented in this paper. An improved gray projection method is presented for inner boundary coarse location. Sobel edge detection and improved Hough transform are adopted for inner boundary fine location. Improved Canny edge detection and Hough transform are used for outer boundary location.2)The lower eyelid is modeled by two parabolas using the Least Squared Criterion and the direction of the parabola is not considered in this dissertation in order to speed up the algorithm. Experimental results show that the proposed method can estimate the eyelids contour with good accuracy.3)Research traditional normalization algorithm published by Chinese Academy of Sciencer and two approach of eliminating the illumination influence. 4)The LogGabor wavelet's performance is analyzed. The parameters of LogGabor filter such as frequency, scale, wavelength, bandwidth, orientation were researched on. Then an approach of extracting feature based on 2D LogGabor filter is proposed. Hamming distance as classifier is also researched.5)In this paper the experiments are done for algorithm validating and simulated iris system. FRR, FAR, CRR are used to estimate the experiments. Plentiful inter-class and intra-calss experiments are carried out to find the optimum values for those parameters in order to provide the best recognition rate. Finally, a simulated iris recognition system is maked to examine the performance of the whole system in the view of practice application.