A novel approach for user authentication using keystroke dynamics

Abstract

One Time Password (OTP) and Text Password are becoming less important in the current age of cybercrime because of the rapid development of new security systems. A user authentication system that is easy, robust, scalable, and cost-effective is a must. For both security and surveillance purposes, keystroke biometrics is a viable option. Behavior biometrics, of which keystroke biometrics is a subset, is used to identify individuals based on the way they type. Typing habits are not set in stone and are subject to change depending on the scenario, the device being used, and the user’s emotional state. As a result, the performance of a keystroke biometrics-based user authentication system is influenced by how well the retrieved information from typing and classification algorithms is processed. Using a keyboard with an array of pressure sensors, this research presents a unique way to keystroke dynamics-based authentication. The goal of this study is to develop user profiles that are unique and different in order to improve the suggested system’s efficiency. A real-world dataset is used to test the suggested method. The outcome is achieved with a 97% success rate in experiments.

Subject Classification:

• 68M25
• 68M18
• 94A62

Keywords:

• Machine learning
• Passwords
• Behavioural biometric
• Sensors
• Keystroke dynamics
• User authentication