AI-Driven Adaptive Authentication for Multi-Modal Biometric Systems

This work covers the application of artificial intelligence in adaptive authentication systems of multi-modal biometric systems. In this paper, a new framework is proposed that employs adaptation in machine learning algorithms for the dynamic adjustment of authentication parameters based on contextual and user-behavior data. This way, all multiple modalities of biometrics, such as fingerprint, facial recognition, and voice patterns, can be utilized for enhanced security and usability. Experimental outcomes show 27% less false acceptance rate and 35% less false rejection rates than traditional static authentication methods. The proposed methodology holds a relative promise toward handling the ever-emerging issues in biometric security with varied environments and user scenarios.