Inter-Device User Keystroke Analysis
Abstract
Keystroke dynamics, the analysis of an individual's typing patterns, has long been studied as a method for user authentication. Key features such as Dwell and Flight Time and Inter-Key Interval have been traditionally employed to distinguish users based on their typing behaviors. While much of this research has focused on desktop or mobile devices separately, analyzing user's typing behavior on both devices together presents unique challenges due to their varied form factors. This paper explores the potential of inter-device keystroke authentication. We use a Siamese Network to identify users across desktop and mobile platforms. By analyzing keystroke data from both types of devices, we investigate the underlying patterns and relationships between them. Our experiments yield high accuracy, reaching up to 98% on a dataset with 37 users and over 73K keystrokes. The results raise important questions about the role of keystroke features in authentication systems.
