Signature Security System for E-Commerce

Abstract

The study of human signatures has a long history, but automatic signature verification is still a new and active topic in the research and application fields of biometrics. This chapter starts with a detailed survey of recent research progress and commercial products in automatic dynamic verification. Instead of applying new and popular approaches, such as those of Artificial Neural Networks, Fuzzy Logic, or the Hidden Markov Model, this chapter proposes a low cost on-line dynamic signature verification method based on the combination of time-dependent global coordinate features and local curvature features. Global features include pen down time, pen down move, the average, maximum and standard deviation of both the velocity and acceleration, while local features make use of the time dependent relationship between adjacent curative turning points. The astonishing growth of the Internet and intranet raises the new challenge of e-commerce security. With the attempts to look for a low cost biometrics method as an enhancement of personal identification in the network, a typical system for dynamic signature verification in the Internet and intranet is introduced. This system involves such processes as dynamic signature data acquisition through the network, using global and local feature extraction, feature match, feature enrollment, and combined feature comparison for verification and distance measures for recognition. Finally, this chapter proposes some applications of on-line signature verification.