Poster: A Continuous and Noninvasive User Authentication System for Google Glass

Abstract

Google Glass stores many kinds of user data, including contacts information, messages and emails, photos and videos, and much more. This brings privacy risks to the owner. To protect user privacy, Google Glass uses four touch gestures selected by the owner from eight pre-defined gestures to authenticate users. This is invasive and also susceptible to peeking. In this work, we propose a continuous and noninvasive authentication system for Google Glass, named GlassGuard. GlassGuard discriminates the owner and an imposter with touch behavioral biometrics and voice features, which are available during a user’s normal daily usage. Touch behavioral biometrics have been studied on smartphones. However, due to user interaction differences between Google Glass and smartphones (e.g. holding smartphones with hand(s) but wearing Google Glass on head, different gestures), systems proposed for smartphones cannot be applied directly on Google Glass.