Hold and Sign: A Novel Behavioral Biometrics for Smartphone User Authentication

Abstract

The search for new authentication methods to replace passwords for modern mobile devices such as smartphones and tablets has attracted a substantial amount of research in recent years. As a result, several new behavioral biometric schemes have been proposed. Most of these schemes, however, are uni-modal. This paper presents a new, bi-modal behavioral biometric solution for user authentication. The proposed mechanism takes into account micro-movements of a phone and movements of the user's finger during writing or signing on the touchscreen. More specifically, it profiles a user based on how he holds the phone and based on the characteristics of the points being pressed on the touchscreen, and not the produced signature image. We have implemented and evaluated our scheme on commercially available smartphones. Using Multilayer Perceptron (MLP) 1-class verifier, we achieved approx. 95% True Acceptance Rate (TAR) with 3.1% False Acceptance Rate (FAR) on a dataset of 30 volunteers. Preliminary results on usability show a positive opinion about our system.