| In the recent years, a new personal identification - biometrics identification system comes into being which aims at the commercial swindle behavior of increasing day by day on Internet at present. The system bases biologic feature such as fingerprint, iris, face and palm line which are exclusive, reliable and steady, and makes use of the powerful function of computer and the technology of the network to carry on image processing and pattern recognition, so it can recognize the identification of human. As a important feature of personal identification, iris has some advantage such as uniqueness, stability, gathering and no-infringing. Comparing with the identification recognition of face and sound, iris is more exact. The system based on iris is composed of iris image acquisition, image preprocessing, feature extraction and matching.In the thesis, the Hough transform is used to extract the centre and radius of circles of inside and outside of iris, thus we can separate the iris from eye. For the sake of remove the displacement and zoom of iris image, iris image is transformed from Cartesian coordinates to polar coordinates and unwrap iris image into a rectangle in the size of 600x100. For improving the effect of iris recognition and reducing the influence of unsymmetrical illumination, the iris image is transformed by hisotgram eaqualization. Wavelet transform is a time-frequency analytical method which is characterized by MultiResolution Analysis. The application of Wavelet transform in image processing is more prominent in all application of Wavelet transform. So we use 2-D Wavelet transform to extract the feature of iris image. Experiment shows that the method Is effective. For iris recognition, We use weighted Euclidean distance. Experiment shows that the method is valuable in iris recognition.
|
PDF Full Text Request