Fingerprint reference point detection using orientation field and curvature measurements

Success of automatic fingerprint identification systems (AFISs) strongly depends on the accurate detection of a reference point on the fingerprints. In this paper we suggest a novel approach for detecting a reference point. First approximate location of singular point(s) is determined by using orientation field and edge detection algorithms. Then to find accurate location of these points, a new secondary filter is introduced. By combining these items the reference point will be obtained. This method is of high accuracy and performance and at the meantime tolerates orientation and translation of the input fingerprint. This can be applied to all classes of fingerprints especially to problematic arch class ones.