A Pattern Matching Classifier for Offline Signature Verification

Offline signature verification can be linked as a biometric classifier that categorizes a signature into two classes: genuine and forged. In this paper a contour matching algorithm is proposed that tracks the basic characteristic patterns in a sample signature and verifies it. It capitalizes on the geometrical properties of the signature and takes into account the inevitable intrapersonal variations for the user set A. The system is trained with 8 original signatures and given a test sample; verification is done by a triangle matching algorithm that validates a signature on the basis of the relative position of the critical points(minimum set of points which when provided can retrace any given signature model). Results verify the effectiveness of the proposed mechanism.