Automated Fingerprint Matching Algorithm And Implementation

Due to the uniqueness and invariability of fingerprint ,automatic fingerprint verification technology became the earliest and most widely used biometric authentication technology. In the past twenty years, automatic fingerprint verification technology got rapidly development and more wide range of application domains. A number of commercial systems exist for fingerprint recognition , however , the performance can't satisfy the rigorous requirements of some special application, and many production wouldn't be publicized. In this paper , based on the research of many literature of automatic fingerprint identification system(AFIS) , two most important technologies——minutiae postprocessing and fingerprint matching algorithms are extensively studied.1.Introduce the developmental history of fingerprint verification system, and the entire work flow.2. Since the fact that many false minutiae raised after the process of extraction, we detailedly discussed on a postprocessing algorithm deleting false minutiae which use local structural information of points. The postprocessing algorithm discussed in this paper can delete most of the false minutiaes and improve the recognition performance.3.Study a subsection feature matching algorithm based on structure minutiae. Aiming at the low recognition performance on condition of poor image quality, use the ridge information of the minutiae to rectify the first matching result. 4.In the second stage of the subsection matching, used changeable bounding boxes to fulfill the fuzzy matching of minutiaes, and to solve the position and angle deviation of the minutiae because of the local deformation of the fingerprint image.5.Study a compound mode multi-condition judge method, which , comparing with the single condition judge method , improve the recognition performance.