Automatic Classifying And Matching Technology Of Fingerprint Image

Biometrics, an automatic personal identification by using some physiological or behavioral characteristics associated with the person, has been accepted greatly with the fast development of electronics, information, computer and networks. As one of the most excellent identifying methods in biometrics, fingerprint identification's theory and application appeal a lot of interest. Automatic Fingerprint Identification System has been researched more deeply and used more widely with many advantages, such as easier to use, higher reliability and lower cost. Fingerprint identification technology has become one of the identity authenticating methods that are the most popular, the most convenient and the most reliable.Although there have been a number of commercial systems and research achievements for fingerprint recognition, but their performances' cannot satisfy the rigorous requirements of some special application. In this paper, the key technology-fingerprint classification and matching algorithm are researched deeply based on the analysis of current fingerprint identification technology reports and dissertations. A new fingerprint-matching algorithm based on region sampling is present:(1) In the section of fingerprint classification, three rotation algorithms are simulated and applied to fingerprint image classification. Based on these algorithms, the question of wrong classification owing to fingerprint image rotation can be resolved efficiently.(2) In fingerprint matching, a new fingerprint matching method based on minutia feature is present. The fingerprint image is partitioned into several regions around the reference point in the proposed method, then the number of minutia points in each region can be counted, and a new feature vector is constructed with them. With this approach, not only all global features of the fingerprint are described (array of the partitions regularly represents the global features), but also local feature cannot be lost (the code of each partition reflect the local feature of fingerprint), so more fingerprint information is used, It can improve the precise of fingerprint identification greatly.All algorithms mentioned in this paper are carried out with Visual C++ on the PIV computer. The experiment results demonstrate that new methods can enhance disposal effects, classification effects and running speed of fingerprint processing.