Privacy-Preserving Biometric Authentication: Fingerprint Encryption and Secure Feature Extraction (pp.349-362)
D. Li1,*, L. Chen1, Y. Zhong1, J. Lian1
https://doi.org/10.30546/1683-6154.25.2.2026.3491Key Laboratory of Intelligent Control Technology for Wuling-Mountain Ecological Agriculture in Hunan Province, Huaihua University,
Huaihua, Hunan 418000, China, e-mails: c2375yu@163.com, zyf325255925@163.com, sakura0503@yeah.net
∗Corresponding author e-mail: ldz1033115125@163.com
Abstract. This study proposes a secure fingerprint authentication framework specifically designed for Medical Cyber-Physical Systems (MCPS) environments to enable clinician authentication for medical equipment such as ventilators, infusion pumps, radiology systems, and remote healthcare infrastructures while ensuring secure and accurate user identification. The proposed framework consists of the following key stages: Data preprocessing, feature extraction, privacypreserving protection using homomorphic encryption and secure multi-party computation, and Secure Hash Algorithm-3 (SHA-3), followed by authentication through a Transformer-based classifier for accurate and efficient fingerprint matching. To further enhance performance, a Hybrid Quantum Swarm Optimization Algorithm (HQSOA) is employed for optimal feature selection and hyperparameter tuning. The system is deployable across device controllers, nurse workstations, hospital edge servers, and cloud telemedicine services, ensuring low-latency, secure communication for repeated clinical workflows. Experimental evaluation on FVC2002, FVC2004, and LivDet 2015 datasets demonstrates superior performance in accuracy, FAR, FRR, and EER, and processing efficiency.
Keywords: Biometric Authentication, Fingerprint Encryption, Homomorphic Encryption, Secure Feature Extraction, Transformer Classifier, Medical Cyber-Physical Systems.
AMS Subject Classification: 62H30, 68T10, 68P25, 94A60.