The Research And Implementation On Essential Algorithm Of Fingerprint Identification System

Currently,traditional security systems which based on token or password cannot meet the demand of modern society. Fingerprint recognition technique which is safer and has more feasibility was appeared under this situation. At present fingerprint recognition technology as a major biometric technology has been widely used in the judiciary, police and other security protection system.Automated Fingerprint Identification System workflow can be divided into: collecting fingerprint image preprocessing, feature extraction, feature matching three parts.the paper focuses on the following content and related improvements:First,using the method based on Gabor filters enhanced fingerprint image, and then use adaptive smoothing threshold of binary images.Finally,using rapid parallel thinning refinement algorithm to get thinning fingerprint image.Extract details of feature points: using feature points of 8 templet to Extract bifurcation point and endpoint. To diferent kinds of false minutiae,this article discussed different methods to detect them and adjusted the methods.Matching of Fingerprints. To raise the shortest distance and the quadtree point model algorithm.The first method of determining reference point based on the shortest distance, then calculated match points based on the reference point, scaling and relative rotation parameters.Finally according to the threshold to determine matching or not. The second method partitioning feature points, and then accordance with the feature points in the right region of the image to judge whether matching or not.