The Research Of Iris Recognition Based On The Wavelet Method

With the development of computer technology especially the Internet technology, a convenient and secure individual identification is required. However, traditional password, PIN and access card all have disadvantages and can't meet the requirement of secrecy .Recently biometric technology has emerged as an ideal solution to this growing problem. Iris is unique and complex organ, which adapts to personal distinguishing feature. Nowadays iris recognition technology is considered as one of the most typical and promising methods in biometric recognition technologies.Based on the iris biometrics ,such processing methods as acquiring, locating, encoding as well as matching for iris image are developed. Such recognition algorithms the feature extraction based on a set of 2-D Gabor wavelet and computation of percentage mismatched bits Hamming distance between a pair of iris representations developed by Daugman, texture analysis methods developed by Wildes and Chinese Academy of Science, extraction of iris signatures by means of zero-crossing of the dyadic wavelet transform developed by Boles and so on. At first we improve the methods of iris image preprocessing, including an efficient arithmetic of evaluation iris images, an improving method of locating iris, calibrating iris image by the translation of coordinate and the circular shift of code, a eatimate of the background illumination by bicubic spline, and the image enhancement by the Dual Tree Complex Wavelet transform. Next we designed the recognition arithmetic based on the location of the local maxima of the wavelet transform modulus.In the thesis lots of simulation are performed by means of MATLAB and Visual C++. The arithmetic of quality evaluation, iris locating, calibration, enhancing and encoding are perfectly realized, and a perfect iris recognition system has been completed. Results of the experiments show that this system has good performance on iris recognition.