Waving Authentication

Abstract

User authentication is important to protect sensitive and private information for smartphone users. We propose Waving Authentication (WA) which is a motion gesture authentication system based on accelerometer. WA utilizes eight distinguishing features hiding in the acceleration traces of motion gestures and exploits one-class Support Vector Machine for classification. It is insusceptible to shoulder surfing attacks. In the interactivity, we first provide two exhibitors' phones for audiences to try intruding WA by all kinds of waving. And we present our WA app to the audience smartphones, letting the phone to recognize their owners on audiences motion gesture.