ENHANCED SCHEME FOR HANDWRITTEN OFFLINE SIGNATURE VERIFICATION

Abstract

Handwritten Signature Verification is a broad area. It has been broadly researched in the last decades but there is an open research problem. There are some possibilities to improve the results. To achieve better results we propose a new handwritten offline signature system. The non-repetitive nature of variation of the signature, because of age and illness this is the limitation of existing system. To overcome this limitation this new system is design. Signature having own psychology or behavior characteristics using that we will propose system to find nature and current psychology of signature person in offline. Different features are extracted from offline signature for comparison. One-class SVM (Support Vector Machine) classifier is used for classification of signature. Signature is genuine or forged one to classify that OCSVM (One-Class Support Vector Machine) used.