A novel biometric system for signature verification based on score level fusion approach

• Published in: Multimedia tools and applications Vol. 81; no. 6; pp. 7817 - 7845
• Main Authors: Dhieb, Thameur, Boubaker, Houcine, Njah, Sourour, Ben Ayed, Mounir, Alimi, Adel M.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2022; Springer Nature B.V
• Subjects: Access control; Biometrics; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Feature extraction; Forgery; Handwritten signature verification; Human performance; Multimedia Information Systems; Signatures; Special Purpose and Application-Based Systems; Sum rules; Support vector machines; Dynamic Time Warping; Biometric system; online signature verification; score level fusion
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-022-12140-7

The active modality of handwriting is broadly related to signature verification in the context of biometric user authentication systems. Signature verification aims to verify a questioned signature as being genuine or forged compared to some previously provided signatures from the claimed person. By doing so, we may be able to verify a person’s identity at accuracy and speed even better than human performance. Application areas of signature verification include different purposes and principally in access controls and forensic document examination. This work presents a novel biometric system for signature verification. We propose a new model that we called the Extended Beta-elliptic model and we integrate the fuzzy elementary perceptual codes (FEPC) to extract static and dynamic features. To discriminate the genuine and forgery signatures of a user, we explore a fusion using the sum rule combiner of three scores which are deep bidirectional long short-term memory (deep BiLSTM), support vector machine (SVM) with Dynamic Time Warping (DTW), and SVM with a new proposed parameter comparator. Our system has been evaluated on two publicly available online signature databases namely SVC2004 Task 2 and MCYT-100, and it shows promising performance gains.