At Your Fingertips: Considering Finger Distinctness in Continuous Touch-Based Authentication for Mobile Devices

Currently, the most prevalent approaches to authenticate smartphones involve either PINs, swipe patterns, or passwords. Few users enable these approaches. In order to encourage adoption, new authentication methods are needed. Emerging methods rely on the distinctness of a user's touch-based gesture for continuous authentication, providing an unobtrusive approach that simply monitors swipes and other input gestures as they are performed in the context of everyday smartphone use. However, existing methods do not consider the distinctness of a user's touch when different fingers are used. In this paper, we present the results of a small pilot study that suggests that a touch-based gesture performed by the same user with a different finger is indeed distinct. We present an approach that uses accelerometer data to identify the position of the phone and the finger that is being used in a touch-based gesture. Our results suggest that touch-based continuous authentication accuracies can be improved by considering accelerometer data and an individual's various fingers.