User Authentication by Typing Pattern for Computer and Computer based devices

Abstract: User access to the most computer systems is secured through possession of a login ID and password combination. Once the login details have been exposed, an unauthorized user has complete access to the computer system. This paper focuses on a powerful authentication system that captures the typing style of a user and detects the change of user based on the typing pattern. The basic idea is to compare a reference set of typing characteristics of the authenticated user with a test set of a user operating the system. The difference between these two sets should be below a certain threshold or else the user change is detected. Moreover on the basis of training sample it has been found that flight time give much better authenticated results when compared with dwell time and words per minute.

Keywords: Dwell time; Flight time; Typing speed; Clustering of keys; Keystroke dynamics;