Off-line Signature Verification based on the Modified Direction Feature

Signature identification and verification has been a topic of interest and importance for many years in the area of biometrics. In this paper we present an effective method to perform off-line signature verification and identification. To commence the process, the signature´s contour is first determined from its binary representation. Unique structural features are subsequently extracted from the signature´s contour through the use of a novel combination of the modified direction feature (MDF) in conjunction with additional distinguishing features to train and test two neural network-based classifiers. A resilient back propagation neural network and a radial basis function neural network were compared. Using a publicly available database of 2106 signatures containing 936 genuine and 1170 forgeries, we obtained a verification rate of 91.12%