AN APPROACH TO DATA COLLECTION IN AN ONLINE SIGNATURE VERIFICATION SYSTEM
Andreea Salinca, Sorin Mircea Rusu, Stefan Stelian Diaconescu · 2012
Online handwritten signature verification is one of the branches of behavioral biometry that is gaining popularity in protecting sensitive information. Our paper addresses a key issue in evaluating performances of online signature authentication systems: data collection. Acquiring a real dataset with handwritten signatures is a major step in the system verification. We will present our collecting techniques in the process of acquisition of dynamic handwritten signatures (more than 5000 genuine signatures and more than 2000 skilled forgeries have been collected from a total of 113 people) useful in the improvement of evaluation results for the authentication system.