The Research Of Iris Recognition Algorithm

With the rapid development of information technology in society, personal authentication is becoming necessary in more and more fields. The traditional personal authentication methods cannot keep up with the requirements of the society because of their inherent defects. Under such circumstance, biometrics emerged as the time requires.As an important composing of biometrics, iris recognition is the most stable and reliable biometric technology. Compared with other biometric features, iris contains more distinctive information. Through analyzing the iris feature and combining with different feature extraction methods, we can realize iris recognition.In paper,include two researches mainly.1. The process of iris recognition includes iris preprocessing, feature extraction and feature matching. In preprocessing, a two-step iris localization approach from coarse to fine is presented in paper according to the distribute of the iris's gray scale. After the iris image is segmented through the k-means clustering algorithm for threshold, the pupil region is isolated and the border of the pupil is located coarsely use the geometric properties of the circle. Then use Gauss filtering to reduce noise and use canny operator to detect inside and outside border, combine the result of rough localization, reduce the blindness of search process, so both pupil border and iris border is extracted fast through Hough transform . Experiments show that this method can extract the boundary precisely and quickly, to meet the requirements of real-time. Because of the difference between people'iris, then transforms the circular iris into a constant rectangular area; use histogram equalization to enhance the normalized iris image.2. The commercial iris recognition systems use 2D Gabor filtering to extract iris texture presently. but it is not a band pass filter in the strict sense, once more than one octave bandwidth, the even-symmetric filter will have non-zero DC component. An iris recognition approach based on 2D Log-Gabor filtering was proposed . Instead of Gabor filters, multichannel Log-Gabor filters were employed to extract the iris phase information. The similarity of two iris codes was measured by their hamming distance. The experimental results in the CASIA database demonstrate that the effectiveness of proposed approach comparing with log-Gabor and 2D Gabor approaches.