New Algorithm for On-line Signature Verification Using Characteristic Hybrid Partitions

On-line signature is a behavioral biometric attribute used for an identity verification. In our previous papers we have proposed methods for the signature verification based on horizontal, vertical and hybrid partitioning. All of them divide signatures into regions called partitions, which are used during training and classification phase. In this paper we propose a new algorithm based on hybrid partitioning, which uses only the most characteristic partitions to evaluate similarity of the test signature to the template. Reduction of the partitions number is used to increase legibility of the verification phase and eliminate partitions which do not increase effectiveness of the verification.