Offline Signature Verification: An Approach Based on Score Level Fusion

In this paper, we propose a new approach for offline signature verification based on score level fusion of distance and orientation features of centroids. The proposed method employs symbolic representation of offline signatures using bi-interval valued feature vector. Distance and orientation features of centroids of offline signatures are used to form bi-interval valued symbolic feature vector for representing signatures. A method of offline signature verification based on the bi-interval valued symbolic representation is presented. Several experiments are conducted on MCYT_ signature database [1] of 2250 signatures to demonstrate the efficacy of the proposed approach based score level fusion for offline signature verification.