Keyboard Invariant Biometric Authentication

Biometric authentication over internet is challenging due to the requirement of additional hardware devices to capture the required biometric data. Keystroke dynamics based approaches may be a good choice for numerous internet applications. However, keystroke data varies significantly across different types of keyboards for a given user (within user variability) which lowers the authentication accuracy, it becomes challenging to use keystroke dynamics in internet applications for authentication. In this paper, we present a unique approach for Keyboard Invariant Biometric Authentication System, which models the biometric data for one type of keyboard using the data from another type of keyboard (laptop). The experimental setup involves a desktop and a laptop computer in heterogeneous environment. Total 98 subjects have been participated in this research. We have compared the flight time of different users using desktop and laptop with the help of key pair. Regression technique is used to map the data for the two types of keyboard.