AI-Driven Dual Biometric Authentication Using Face and Voice Recognition for Secure Smart Homes

Smart homes require robust security systems to mitigate threats posed by unauthorized access. This paper proposes a dual biometric authentication system combining facial and voice recognition, empowered by deep learning models and enhanced with liveness detection and privacy-preserving algorithms. The proposed approach integrates a weighted fusion strategy, bone-conduction-based liveness detection, and homomorphic encryption. The framework is validated using simulated data and achieves an authentication accuracy of 98.3%, significantly outperforming unimodal systems.