Improving Performance and Usability in Mobile Keystroke Dynamic Biometric Authentication

In the last few years, the number of mobile devices such as smartphones and tablets, in circulation, has increased dramatically. The primary and often only protection mechanism in these devices is authentication using a password or a Personal Identification Number (PIN). Passwords are notoriously known to be a weak authentication mechanism, no matter how complex the underlying format is. A more secure alternative option which has gained interest recently is extracting keystroke dynamic biometrics from supplied passwords for mobile authentication. In this paper, we show that using random forests classifier, improved accuracy performance can be achieved for mobile keystroke dynamic biometric authentication. We also propose a new algorithm for handling typos, which is an essential step in improving usability. We study both timing features and pressure-based features. Experimental evaluation is based on two public datasets and a third dataset collected in our lab. The best performance, obtained by combining timing and pressure features, is an Equal Error Rate (EER) of 2.3% for a population of 42 users.