Research Of Iris Recognition Based On Entropy And Moment Invariants

With the rapid development of science and technology, human's consciousness of security is improved gradually. The traditional security technology doesn't meet the requirements of security quality at present. Then, the people turn to biology recognition technology and make it develop fast. Among the different kinds of biological recognition techniques, iris recognition technique is studied in depth and applied widely for its advantages, such as reliability, uniqueness and non-forgery.In this paper, some key issues of iris recognition technology have been deeply studied .The main works as follows:First of all, an iris location algorithm based on the combination threshold method with Canny operator has been proposed. The steps of the algorithm are implanted in order. Pupil is separated from eye gray image by the threshold segmentation method. Then the inner edge of the iris is found out with the border tracking method. The edge of the iris is obtained by combining the Canny operator with the classic Hough transform. So as to the region of iris is located.Secondly, a method by means of eliminating the big connect region is represented in order to clear the eyelid and eyelash from the iris image. The means not only preserve the iris features, but also come up to eliminating the noises from iris images. Thirdly, an algorithm of extracting iris feature and matching iris is put forward, which is based on entropy and the same moments. In the algorithm, the iris image is broken into some small blocks firstly, and the entropies of each block are calculated and these entropies form an image entropy matrix. Then ten invariant moments of the matrix are extracted to be the eigenvector in process of iris recognition. Meanwhile, these moments are saved into database as a template of iris recognition. The Euclidean distance between two eigenvectors is the criterion for the iris recognition.A lot of experiments have been carried out to test the proposed algorithms. And the experimental data have been analyzed in details. The results showed that the same person from the iris eigenvectors Euclidean distance in 99% of the data (0.7 ~ 0.9) , and different characteristics of the iris 99% of the distance vector data (1.2 ~ 1.6) range.. The experimental results also show that the algorithm has rotation invariance and scaling invariance.Meanwhile, the algorithm has some advantages, such as stored data less, matching more rapidly, match result more effective.