Gait-Watch: A Gait-based context-aware authentication system for smart watch via sparse coding

Abstract In recent years, wrist-worn smart devices such as smart wrist band and smart watch have pervaded our everyday life. Under this trend, the security issue of these wearable devices has received considerable attention as these devices usually store various private information. Conventional methods, however, do not provide a good user experience because they either depend on a secret PIN number input or require an explicit user authentication process. In this paper, we present Gait-watch, a context-aware authentication system for smart watch based on gait recognition. We address the problem of recognizing the user under various walking activities (e.g., walking normally, walking upstairs and walking with calling the phone), and propose a feature extraction method from gait signals to improve recognition accuracy. Extensive evaluations show that Gait-watchimproves recognition accuracy by up to 30.2% by leveraging the activity information, and can achieve 3.5% Equal Error Rate (EER). We also report a user study to demonstrate that Gait-watchcan accurately authenticate the user in real-world scenarios and require low system cost.