Real-Time Multi-Factor Biometric Encryption for Secure and Scalable Authentication Systems
K. Gowtham, A. Krishnaveni · 2025
The growing threat of biometric data theft and identity limitations has contributed to the demand for better and privacy-preserving authentication mechanisms. This paper puts forward a strong multi-factor biometric encryption framework with a combination of fingerprint, iris, and password-based modalities to promote secure authentication. The framework provides integrity and confidentiality of the template through an algorithm utilizing advanced feature extraction techniques – Gabor filters in the case of fingerprint and SIFT in the case of iris, AES-256 encryption, and SHA-256 hashing. Principal Component Analysis (PCA) can be processed in real-time by feature reduction, and privacy-preserving matching is possible using Cosine Similarity with TenSEAL. The multimodal dataset under consideration can give the authentication accuracy of 99.20%, False Acceptance Rate (FAR) of 1.70%, False Rejection Rate (FRR) of 2.30% and an Equality Error Rate (EER) of 1.90% based on the experimental analysis. The proposed system also provides an acceptable trade-off between security, real-time, and scalability over the most related models of state-of-the-art, than the other new state-of-the-art models of multimodal biometrics encryption. It is a cloud/edge deployable architecture that meets ISO/IEC 24745 standards of biometric template security. The system is scalable with modular integration in high-security systems like digital ID, banking, and military-grade access control systems. This research work will bring both academic excellence and practical biometric security through the integration of physiology and knowledge-based authentication into encrypted domain processing.