Occluded Fingerprint Recognition Algorithm Based on Multi Association Features Match

Abstract

Recognition rate of mainstream fingerprint recognition algorithm is very low for occluded fingerprint image. In order to solve this problem, based on multi association matching features this thesis proposes the recognition algorithm (RA-MAMF). Firstly, image is pre-processing, the fingerprint image's Gabor filter is enhanced, and bi-narized and thinning pretreated are involved; secondly, the image is divided into multi homogeneous subsets, in which statistical association features and the bifurcation points of each subset are respectively extracted; finally on the basis of fingerprint, images are to be recognized, they are compared and match with the subset. Complete and occluded fingerprints data sets are used to make tested; the recognition algorithm based on multi association matching features has achieved excellent recognition accuracy; the RA-MAMF algorithm does not significantly increase the operating time, and this method effectively solves the low accuracy of traditional identifying the Occluded fingerprint image.