| Recently, biological recognition technology as a kind of identity recognition method is more and more emphasized. Biological recognition technology is a kind of pattern recognition based on biological technology, which makes use of information technology and incorporates the biology and information technology. It is a kind of technology to recognize the identity according to the human physiological and behavioural characteristic. Because the biological characteristic is unique and steady during a period, and it is not easy to be faked and forged, the biology recognition technology is used to recognize identity, and it is safe, reliable, and accurate. Moreover, the biological recognition product aided by the computer technology is easy to implement the functions including safety certification, monitoring and management. The breakthrough of biological recognition technology will not only bring the great development of the theory of safety certification, monitoring, management and pattern recognition, but also bring great social and economy effect.Because the fingerprint and iris recognition has the features including: it is easy to acquire, its feature is apparent, it is unique and it is not easy to fake. Moreover, the fingerprint recognition is easy for people to accept, and the iris recognition has high accuracy and it is long-time steady and difficult to fake, so they are the generally accepted promising biological recognition technology, and they have important theoretical and social economy meaning. We make research of the fingerprint and iris recognition technology, and their breakthrough is certain to bring great benefit for human being.In this dissertation, by the research of recognition technology of fingerprint and iris, we mainly make some improvement and innovation in the follows fields: fingerprint enhancement, fingerprint classification, fingerprint match, iris auto-acquisition, iris location, feature extraction and iris match. Generally, the fingerprint is easy to be polluted and affected, consequently, the quality of fingerprint images will be lessened. Under nonideal acquisition condition, we have made improvement and recovery processing from fingerprint images. The direction-based adjustment method of fingerprint enhancement is introduced, which is more accurate to indicate the real texture information and the computation speed is greatly increased. At the same time, based on the information extraction method, Gabor filter-based image enhancement method is put forward. By handling of low quality fingerprint images, the accuracy of feature extraction is noticeablely improved. The speed of the auto fingerprint recognition system can be increased by the fingerprint classification algorithm, on the contrary, the accuracy ratio of existed algorithms is usually low. Aimed to the problems in the four noteworthy aspects in auto fingerprint classification, the traditional Core point and Delta point model is researched, and its definition method is developed. A kind of fingerprints classification algorithm based on the new definition of the Core point and Delta point, and its experiment data have the higher classification accuracy.The fingerprint match is implemented by matching the feature pattern in two given fingerprint images, and it is judged that whether the two images is from the same finer. The computation of existed fingerprint match algorithm is great, and they can't satisfy the need of real-time recognition, so it affects the application of auto fingerprint recognition system. In order to solve the problem, according to the shape structure of fingerprint, we put forward three kinds of basic fingerprint match methods of fingerprint images and structure feature: 1. The fingerprint match algorithm based on curve structure and triangle method. By the tracing sampling of special feature point in fingerprint images—the connected curve between isolate points in the center part of an image, the two fixed points of the curve are got, then the vector triangle method is used, the reference point...
|
PDF Full Text Request