Abstract
Authentication is a cumbersome task for users, as it needs to be repeated multiple times a day. While the introduction of fingerprint authentication on mobile devices reduces users’ memorization and typing efforts, users still need to perform a specific action to unlock, e.g., their smartphone. In order to further reduce the associated overheads, we analyze in this paper the feasibility of identifying users based on the way they pick up their device located on, e.g., a table. By doing so, no dedicated actions would be required from the users, as the authentication motion is already done when they want to access different services available on their devices. We have therefore designed and conducted a lab study involving 24 participants to collect the sensor data associated to these pick-up motions. The obtained classification results are promising and suggest that motion sensors could be integrated in a future implicit authentication scheme.