Fingerprint singular points detection and direction estimation with a "T" shape model

Abstract

As a sort of evident landmark features of fingerprints, singular points (SPs) play important roles in fingerprint alignment, classification and recognition. We present an adaptive "T" shape model of SPs and develop a robust and generic approach to detect SPs and their directions simultaneously. The proposed approach utilizes homocentric sectors around candidate SPs to pick out lateral-axes and further main-axes based on the proposed model. The results of the experiment conducted on a public database, FVC 2002, demonstrate the effectiveness of the method in this paper. © Springer-Verlag Berlin Heidelberg 2005.