An Ensemble Model for User Authentication Leveraging Keystroke Dynamics and Facial Recognition
Taduvai Satvik Gupta, Konduru Praveen Karthik, Sri Sai Suhas Sanisetty, Sagar Basavaraju · 2024
User authentication is a process that confirms a user's identity before allowing entry into a system. This paper presents a multi-modal user authentication system by integrating keystroke dynamics and facial biometrics of a person. The system uses machine learning techniques for keystroke recognition and deep learning for facial recognition techniques, for building a secure user authentication system. In addition to this, the model also alerts the user if there is any frequent unauthenticated user trying to access their account. From the experimental results it has been noted that the model is achieving an accuracy of 93.4 % for the keystroke detection by giving optimal results compared to the Benchmark dataset and the face recognition model is also achieving an accuracy of 96 % with slight increase in the performance compared to the SOTA models. By leveraging multiple biometric modalities, the proposed ensemble model achieves higher accuracy and resilience against spoofing attacks, ensuring secure user authentication.