Keystroke Rhythm Analysis Based on Dynamics of Fingertips

Abstract

The proposed work presents an analysis of rhythmic patterns based on the dynamics of the fingertips observed during keystroke events on a traditional computer keyboard. In this work, a detailed analysis of pressure and acceleration applied by the user has been taken into consideration, in contrast to the earlier works which have focused primarily on the timing variations of the keys. In this purpose, two different types of sensors pressure and accelerometer were embedded on a traditional keyboard. Groupings of numerical digits and special keys were used to design different kinds of tasks to acquire sensor and timing variations of each keystroke event. In total, six subjects (two females and four males) were asked to provide the data to validate the proposed idea. Subsequently, two types of analysis, interpersonal and intrapersonal keystroke rhythm, have been carried out. The analysis results show uniquely identifiable keystroke sensor and timing variations for different users. However, individual users almost maintained his/her keystroke pressure and acceleration variation irrespective of the type of tasks. Such results demonstrate that keystroke rhythm based on pressure and acceleration variations of the fingertips can be used as a behavioural feature for developing more sophisticated biometric systems for intrusion detection.