In the offline world, payments are often made over the counter with some levels of human inspections. However, such inspection does not exist in the online world. Online transactions are carried out virtually on a remote application server. Though other technical security measures have been introduced for online transactions such as the use of encryption, digital signatures and digital certificates, the issue of trust is still largely a major problem in e-commerce since actual authentication of users is not often established. A solution to this is to use biometrics Automatic Signature Verification (ASV) systems where human identification is carried out automatically based on their signatures. The main advantage of ASV over other biometrics technologies is that its applications are widely accepted and generally acknowledged by the public due to the fact that signatures have long been used as proof of identity in legal documents and financial transactions. Additionally, the ASV system allows the extraction of dynamic information that describes the way a signature is actually executed in terms of velocity, acceleration, pen pressure, pen inclination, etc. Many signature experts believe the dynamic information of the signing operation is generally consistent and stable throughout one’s lifetime. This in turn is more secure simply because it is harder to imitate human signing operation than to reproduce signature images of another person. Since ASV allows for remote networked authentication, it appears promising to most e-commerce applications. This paper generally describes the ASV potentials, its current applications and impediments in e-commerce related activities. It also addresses areas for ASV improvements.
Keywords: Automatic Signature Verification, biometrics, FAR, FRR, EER, feature sets, feature weights, feature thresholds.
ICoEC Online Proceedings Library