Off-line signature verification using global & local features with neural networks

In 21st century, along with extravagant diffusion of Internet, multimedia technology and a growing need for individual verification in many day to day applications, automatic signature verification is perused with more interest. Signature verification plays a vital role in a enormous number of fields starting from passport verification systems, online banking, to even authenticating candidates in various public examinations from their signatures. Even today many of business and commercial transactions are being authorized by means of signatures. So, an automatic signature verification system is required. This paper presents an Off-line Signature Verification system (OSVS) where the strong feature set thus obtained makes the OSVS more accurate. The effectiveness of proposed feature set has been investigated over 310 signatures using a neural network classifier. The performance of the proposed system is evaluated by calculating False Rejection Rate (5.0%) & False Acceptance Rate (5.8%).