Learning a Policy for Gesture-Based Active Multi-touch Authentication Raquel Torres Peralta, Anton Rebguns, Ian R. Fasel, and Kobus Barnard Universidad de Sonora, Departamento de Ingenieria Industrial, The University of Arizona, Department of Computer Science, Tucson, AZ 85721-0077 {rtorres,anton,kobus}@cs.arizona.edu, ianfasel@gmail.com Abstract. Multi-touch tablets can offer a large, collaborative space where sev- eral users can work on a task at the same time. However, the lack of privacy in these situations makes standard password-based authentication easily compro- mised. This work presents a new gesture-based authentication system based on users’ unique signature of touch motion when drawing a combination of one- stroke gestures following two different policies, one fixed for all users and the other selected by a model of control to maximize the expected long-term infor- mation gain. The system is able to achieve high user recognition accuracy with relatively few gestures, demonstrating that human touch patterns have a distinc- tive “signature” that can be used as a powerful biometric measure for user recog- nition and personalization. 1 Introduction Tabletop devices allow the interaction by touch using a comfortable interface highly visible to those individuals close to the display. This particular characteristic has made the authentication of a user a challenging task. Currently, one solution is to use a password. However, a relevant issue with touch devices is the lack of privacy while typing characters. The same problem applies to a gesture-based authentication systems. The alternative then is to use special “signature” a user leaves unconsciously on the touch of the surface and within the gesture itself. But, does this special signature exists? To answer this question, we examined basic one- stroke gestures of 8 different users using different representations based on speed and shape with the goal of finding the features that differentiate one user from the rest when drawing a set of gestures.In this paper, we propose a new gesture-based authentication system for touch devices, which does not consist on a secret combination, but rather in the shape and speed when drawing a particular gesture. The authentication can be made after a series of gestures requested systematically and strategically by the system with no need of privacy. Although no high-precision equipment was used, it was possible to achieve high recognition rates after a few gestures. L. Marinos and I. Askoxylakis (Eds.): HAS/HCII 2013, LNCS 8030, pp. 59–68, 2013. c  Springer-Verlag Berlin Heidelberg 2013