Biometric Evolution: Leveraging Semi-Supervised Learning for Secure Identification

In recent years, cybersecurity has elevated to the maximum level of interests for governments worldwide. Ensuring robust cybersecurity is essential for safeguarding computer systems, networks, and data from cyberattacks, which can have detrimental effects on individuals, businesses, and governmental operations. Biometrics has emerged as a crucial component in cybersecurity, effectively preventing unauthorized access, identity theft, and unauthorized data modifications. This paper introduces an innovative semi-supervised deep rule-based classifier designed for cybersecurity applications. Our methodology integrates deep features extracted through advanced deep learning-based image analysis techniques, including DCTNet, DSTNet, PCANet, and ICANet. The semi-supervised deep rule-based classifier employs both labeled and unlabeled data to enhance classification performance. Experimental results, obtained using a standard database, demonstrate exceptional identification rates, significantly surpassing those reported in previous studies.