On the Effect of Typing Hand on Mobile-based Keystroke Dynamics

With the proliferation of mobile devices, access control enabled by biometrics is going to be an indispensable function. In this study, we consider mobile biometrics based on touchscreen based keystrokes. Although the general topic of keystroke dynamics have been well studied, touchscreen based keystrokes remains challenging and relatively new. Few studies evaluate the effect of typing hand on mobile biometrics. We therefore set out to formalize the problem and explore a family of solutions based on keystrokes and typing hand, which can be left hand, right hand, or both hand. We address open questions such as: the effect of typing hand on touchscreen keystrokes, whether or not the knowledge of typing hand is important, and the effect of user-specific score normalisation on the performance. General Terms Security, Privacy, Algorithms, Machine Learning