Research On The Automatic Fingerprint Identification Technology

With the development of the society and economy, the accuracy, security and practice of the personal identification are highly required. Due to the hidden-troubles of being lost, forgotten , copied and stolen easily of traditional personal identification of token or password, biometrics that does the identify authentication by identifying the personal physiological and behavior traits provides a convinent and reliable solution . Based on the ecological traits, with a method of information-processing technology, biometrics combines the ecological technology with information technology organically. Among the numerous biometrics, more attentions have been paid to the fingerprint identification technology due to its convinence, high accuracy and low cost. The fingerprint identification, mainly used in electronic commerce, criminal identification, information security and etc, has become the most effective method for personal identification.Based on the achievements of Automatic Fingerprint Identification Technology selected and analysed at home and abroad in recent years, this paper aims to make studies on algorithms of AFIS. Mainly works are as following:(1) The segmentation of fingerprint image. Low-quality fingerprint image processing is the focus of research on the automated fingerprint identification technology in recent years while the low-quality fingerprint image segmentation is the prerequisite for achieving the follow-up treatment. In the paper, based on the analysis of variance as a separate indicator of the limitations ,we sutudies the gray-scale distribution of fingerprint image, and advances the theory of fingerprint image segmentation algorithm based on the texture. The results showed that, compared to the variance based on the gray fingerprint image segmentation algorithm, the algorithm of the division in the essay is much better, and the noise of the resistanceis much stronger.(2) The extraction of fingerprint feature. This paper studied two commonly used 8 - Cross neighborhood of algorithms. Extracted all of the client nodes and branches from the fingerprints, and on the basis of this proposed a new 8 - Cross neighborhood of algorithms. The experiment proved that the algorithm can improve the accuracy of extractionon feature points.(3) Probed deeply into elaborating a way to fingerprint image thinning image pre-processing methods such as fingerprinting technology, concrete realization of the difference by comparison, statistics from the poor, gradient method to strike a fingerprint pattern, with rapid thinning algorithm to improve The OPTA algorithm, based on the 4-connected parallel thinning algorithm serial algorithm refinement fingerprint image.(4) Probed deeply into studying the fingerprint matching algorithm based on similar principles of the triangle pattern-matching algorithms and point on the ridge line calibration,and elaborating the characteristics and advantages of the two methods. Proposed the key follow-up studies of the fingerprint matching technology through a comparative analysis of their respective advantages and disadvantages...