Online Handwritten Signature Verification Discriminators Based on Global Feature

Contemporarily, the internet has been heavily used for the electronic commerce especially in the areas of finance and banking. The transactions of the finance and banking on the internet involve use of handwritten signature as a symbol for consent and authorization. Handwritten signature is one of the biometric techniques that are widely accepted as personal attribute for identity verification. Hence, it is vital to have an online handwritten signature verification system that is fast, reliable and accurate to avoid attempts to forge handwritten signatures, which has resulted in heavy losses for various financial institutions. This paper presents the implementation of an online handwritten signature verification system (OHSV) using dynamic features as the discriminators. It will describe the functions and modules of the system, explain on the approach used, and discuss the performance results of the system, which are measured based on the false rejection rate (FRR), and false acceptance rate (FAR). The former means the rate of genuine signatures that are being incorrectly rejected while the latter means that forgeries that are incorrectly accepted. The experimental results showed that the features based on number of stroke and vertical speed are sufficient to discriminate genuine samples from forgery sample based on the given threshold.