Iris Recognition Algorithm

With the development of information technology in contemporary society, the demands for security are becoming more and more important. So the technology of biometrics identification has been improved and applied in many fields. As one of biometrics technology, iris identification has been gradually applied to safety fields. Iris identification is based on iris texture to identify persons, and it is one of most accurate biometrics technologies.The whole iris identification system is composed of iris image acquiring, image preprocessing, feature extracting and matching. Based on iris images database provided by Institute of Automation Chinese Academy of Science, each part of iris recognition has been thoroughly studied and discussed in this thesis. This paper presents some new improved algorithms which gets a better result in the experiment. The main work of the dissertation is as follows.1. In iris localization, the algorithms of Canny edge detection and Hough transform are improved to detect the outer and inner.boundaries of iris, which reduces the boundaries for researching and enhances localization speed.2. For the effects of eyelid and eyelash occlusion, a new method of iris noise eliminating is presented based on Radon transform, which can effectively eliminate the effects for eyelid noise information. For eyelashes eliminating, the threshold technique is used. Experimental results show that the method is effective in restraining iris noise and improving iris recognition rate.3. In the image normalization, the localized iris is transformed from a circinal area to a rectangular area, then the normalized image is enhanced.4. Conventional iris feature extracting and encoding methods are discussed and analyzed in detail. Three improved algorithms are given: (1) an iris recognition algorithm based on multichannel wavelet filter; (2) an iris feature extraction algorithm based on 2-D wavelet transform and directional vector; (3) an iris recognition algorithm based on local walsh transform.5. In methods of pattern matching and classifier designing, an improved Hamming distance classifier and a weighed Euclid distance classifier are applied for pattern matching according to different algorithms. At the same time, in order to reduce angle rotation of iris, we work out a method based on circulation shifts, which effectively reduces the effects of angle rotation and ulteriorly enhances iris recognition rate.Experiment results show the effectiveness of these algorithms, and a higher accuracy and speed are achieved. It is proved that the iris recognition arithmetics proposed in this paper can be performed as a good arithmetic foundation for an identification system.