Study And Implement Of Iris Recognition Algorithms

With the development of information technology and the increasing emphasis on security,automated personal identification based on biometrics has a rapid development in recent years.Iris recognition technology with its huge potential is relatively new among all kinds ofbiometric recognition technologies. It shows great advantages. Firstly, Iris texture possesses ahigh level of randomness in its patterns. Even the irises from different persons with the samegenotype are different. Secondly, the characteristics of iris are steady in a whole life. Thirdly,the capturing of iris image is non-invasive, which is very convenient for the users of the irisrecognition systems. Fourthly, iris tissue possesses the characters of living tissue because ofthe zoom of pupils, which can prevent counterfeiting.In this dissertation, biometric technology and iris recognition technology have beenintroduced. Then, the structure of iris and component of iris recognition system have beendiscussed in detail. Finally, the algorithms of iris recognition have been studied. Algorithmsproposed in this dissertation have been simulated on CASIA iris image database by usingMATLAB language. The main work includes:1. Iris location algorithm has been studied. Iris location algorithm based on Canny operatorand Hough transform has been implemented.2. Iris normalization algorithm has been studied. Iris normalization has been achieved byanti-clockwise mapping the iris to a rectangular block of a fixed size.3. Iris enhancement algorithm has been studied. Iris images have been enhanced by means offiltering the image with a low-pass Gaussian filter and local histogram equalization.4. Principal component analysis algorithm has been studied. Iris recognition algorithm basedon principal component analysis is proposed in this dissertation.5. Independent component analysis algorithm has been studied. Iris recognition algorithmbased on principal component analysis and independent component analysis is proposed inthis dissertation.6. Two-dimensional principal component analysis algorithm and its improved algorithm havebeen studied. Iris recognition algorithm based on Two-dimensional principal component analysis and iris recognition algorithm based on modular two-dimensional principalcomponent analysis are proposed separately in this dissertation.