Study On Iris Identification Algorithm

With the development of the information age, information security has become moreand more prominent in social life. Biometric identification technology can meet thecurrent and future needs of the development of identification technology, which constantlymake its advantages be highlighted and usher it a broader space for development.Depending on the superiority, iris recognition technology is more in keeping with thedemands of peoples that are safe, accurate and stable, on identification technology, amongthe existing technologies. And it naturally became one of the most promising biometrictechnologies.Firstly, this article briefly describes iris recognition technology, including itstechnical characteristics, development status, development prospects, and the basictheoretical knowledge.Secondly, starting from the composition and structure of iris recognition system, thispaper conducted a study on algorithm for it, containing iris location, iris normalization,iris image enhancement, iris feature extraction and matching. In iris location the previousalgorithms are studied and analyzed, and then an iris location algorithm is proposed basedon a combination of gray-scale characteristics of the boundary points and geometricprinciples. This location algorithm is a simple calculation method, it have high speed andaccuracy. In iris feature extraction, several previous classic algorithms are discussed andstudied. The study centered on the iris feature extraction method based on PCNN model,and proposed an improved method for the deficiencies in the previous algorithm. Afteranalyzed in the experiment, the proposed algorithm is proved to be effective in extractingthe characteristics of the iris texture information. In the iris feature matching, the hammingdistance matching method is chosen to match the iris feature, with the iris code shiftedcircularly during the matching procedure. And the appropriate shift number for this articleis settled by the experiment to improve the match rate.Finally, the studies on each step of the iris recognition system in this paper were allsimulated. The results show that the accuracy and speed of the proposed location algorithm are both improved. The feature extraction method is fast and effective, and itreduced the data size of the iris code. Finally, the proposed algorithm gained the idealequal error rate and recognition rate in the match step.