Font Size: a A A

Study And Implementation Of Identification Recognition Algorithms Based On Iris

Posted on:2007-04-24Degree:MasterType:Thesis
Country:ChinaCandidate:W Z GaoFull Text:PDF
GTID:2178360182473758Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Utilizing some biological characteristics of human body, which are unchanged and unique in all one's life,is an important research direction in the recognition technology at present. In this paper, the physiological structure and characteristic of iris are introduced. And then the characteristic of iris recognition is also expressed. The development of iris recognition technology and the meaning are discussed briefly. The algorithms of iris recognition are researched deeply and all the processes that include location, normalization, feature extraction and pattern match are carried out.In the iris location respect, to the characteristic of the iris image of this database, two methods are adopted: Canny combines with Hough transform method and the circle template method. The former at first utilizes Canny to detect the edge of the iris image, uses the Hough transform to locate the interior circle and the other circle; The latter utilizes gray projection and the threshold separate to find out the interior circle, utilizes the circle template operator, which has been put forward by Dr. Daugman, to determine the other circle. The experimental results show that Canny combines with Hough transform method can locate the iris images more effectively. The normalization of the iris image realizes the procession that the iris varies from circle to the rectangle, making the iris do not have the elasticity and translation. Meanwhile, the method of utilizing the histogram equalization dispels the influence because of the non-uniform illumination. For feature extraction, two kinds of algorithms are adopted. They are the coefficient correlation estimates encoding and the Gabor transform encoding. And the selection of child images that influences the results of iris recognition is studied in detail. Hamming distance is applied to the pattern match and the consequences analysis. And the result is very good. At last, the results and conclusions of the paper are pointed out. At the same time, some expectations to the future research are predicted.
Keywords/Search Tags:iris recognition, iris location, normalization, feature extraction, Hamming distance
PDF Full Text Request
Related items