Research On Iris Recognition Algorithm

In modern times, going with the development of science, technology and economy, people think much of identity recognition more and more. But traditional identity recognition technologies such as password etc. have not suit the identity recognition in modern times for their bad ability against forging. For overcoming the disadvantages of traditional identity recognition technologies, biometrics recognition methods were brought forward such as fingerprint, iris and voice. Among these biometrics recognition methods, the research on iris recognition has received increasing attention because of its high reliability, In this dissertation the iris recognition algorithm was researched, and the main works are as follows:Firstly, researched the location of the iris inner boundary, and put forward two iris inner boundary location methods. The first method fixed on the gray scale threshold for pupil's segmentation by finding the pupil's gray scale range, based on the gray scale histogram of the iris image, and located the pupil by gray scale projection. The second method researched deeply the characteristics of iris image's row and column gray scale curves, and put forward locating the iris inner boundary by the characteristics. The two methods can all locate the iris inner boundary successfully.Secondly, the edge detection methods were researched based on human's visual characteristics for designing simple and reliable edge detection method to locating the iris outer boundary. And put forward several edge detection methods based on human's visual characteristics: the method of the gray scale ratio, the negative gray scale ratio, the gray scale difference-sum ratio, the negative gray scale difference-sum ratio, the gray scale proportion logarithm difference, the negative gray scale proportion logarithm difference, the gray scale integer proportion logarithm difference, the negative gray scale integer proportion logarithm difference and the gray scale proportion power difference. The test results show these methods can all detect image edges effectively.