Enhancing signature verification using alternative handwriting semantics

This paper briefly reviews the context of automatic signature verification and addresses the issue of using alternative handwriting semantics as an additional source of information for authentication purposes. As a foundation for this study, two simple signature verification systems are developed. The area of biometric fusion is reviewed and some approaches are proposed for combining several different experts within the handwriting modality. Thus, signatures are fused with handwritten PIN-codes in order to evaluate whether extra information extracted from alternative handwriting semantics can lead to better overall system performance. We explore how to develop a handwriting recognition system which can evaluate the best possible digit combinations to constitute user-specific PIN-code choices, based on personal handwriting characteristics. Finally, an alternative strategy which involves handwritten passwords is briefly considered. Overall, therefore, the work provides some new insights and objective practical data to support further evaluation of enhanced signature verification systems for identity checking applications.