Adaptive Continuous Authentication System Based on the User's Emotional and Contextual State
This article addresses the problem of low accuracy in continuous authentication systems caused by the natural variability of user behavior. An analysis of existing biometric approaches is conducted, justifying the selection of an adaptive two-stage model as an effective method for accounting for the user's psycho-emotional state. The authors designed the AURA (Automatic User Recognition Agent) system using a component-based approach, which allowed for a clear separation of the state identification and authentication tasks.