Intelligent biometric signature verification system incorporating neural network

Abstract

In this paper, the practical development of a novel but yet simple, low cost and userfriendly PC-based biometric technique to verify human signature is presented. The system is based on data acquired from time duration, pen position and pressure signals of handwritten signatures. A preprocessing stage of the acquired signature signals involving the use of a number of common feature extraction techniques is rigorously highlighted. An intelligent feature of the system is made possible through the application of a multilayer feedforward neural network that is used together with suitable algorithms to complement the verification process. The results of the study showed that the system is effective and promising in identifying correct human signatures presented to the system.