Gabor Transform Algorithms In Iris Recognition

With the development of social information, iris recognition has become one of the most promising biometric technologies. Recently, more and more researchers work for the iris recognition and some algorithms are presented, but these algorithms are sensitive to the noise and bring negative influence to the iris recognition system.A series of algorithms based on the analysis of the texture feature are presented to pretreat the iris image and extract characteristic from the iris image by analyzing the problem existed in the iris recognition system and studying Gabor Transform. First, a2-D Circle Gabor Filter is designed to estimate the position of the pupil and detect the specular reflection, then a1-D peak Gabor Filter is proposed to detect eyelid boundary and eyelashes. Furthermore, eyelid is localized on eyelid boundary image by parabolic Integrodifferential operator in order to avoid the impact of the double-fold and the root of the eyelashes. Finally, a Gabor coding filter is designed to extract peak and gradient information from the iris region.The result which experimented on the data CASIA-IrisV3-Lamp showed that the presented algorithm can avoid the impact of the hair and the big black speck on estimating the position of the pupil, remove the noise information effectively. The correct rates of estimating the position of the pupil and the parabola localization are100%and99.5%respectively, meanwhile, the proposed algorithms lead the accuracy of pupil localization and iris localization to increase by9.8%and16.7%respectably, and the correct rate of the iris recognition is99.78%. The experimental results show that the proposed algorithms have better stability and versatility.