HANDWRITING FOR PEN COMPUTERS ON-LINE SIGNATURE VERIFICATION

Abstract

The handwritten signature is commonly used to authenticate a financial transaction or manifest the contents of a document. Signature verification is usually done by visual inspection. However, automating the signature verification process will improve the situations where a signature is required and eliminate falsification. A wide range of methods for online signature verification has been proposed in the literature. On the other hand, the so-called Support Vector Machine (SVM) has been recognized as a powerful technique for pattern classification. In this paper, we developed an on-line signature verification system based on SVM. The system is then tested using a data set of sample signatures, which includes genuine and forgery signatures. The error rates obtained by applying the system on the proposed data set were promising and encouraging its application in the industry.