Waving Gesture Analysis for User Authentication in the Mobile Environment

The increasing popularity of wearable devices has brought great convenience to human life and business. As wearable devices have become widely used personal computing platforms, more and more private information gets accessed by them, which stresses an urgent need for feasible and reliable authentication mechanisms in the current mobile computing environment. However, traditional memory- based authentication methods like PINs have been proven easy to crack or steal. Based on the fact that hand-waving patterns vary among different users, we propose a novel hand-waving- based unlocking system using smartwatches, which consists of data acquisition, data preprocessing, feature extraction, and authentication modules. Furthermore, we established a 150-person-time hand-waving dataset with a smartwatch, and conducted a systematic performance evaluation, achieving an equal error rate of 4.27 percent in the zero-effort attacking scenario and 14.46 percent in the imitation-attack scenarios. Additional experiments on usability to operation length and sensitivity to sampling frequency are offered to explore the applicability and effectiveness.