User authentication method using shaking actions in mobile devices

Abstract

In order to use mobile devices safely, many researchers have worked on topics related to mobile security like user authentication. This paper proposes a user authentication method using users' biometric information. The propose system uses the device movement data that are obtained when a user shakes his or her mobile device. To capture the biometric information (shaking actions), acceleration data and rotation angle values of a mobile device are measured by the proposed system, and noises are removed in the preprocessing phase. After the preprocessing phase, the system extracts features and classifies the users based on these features. We had experiments to measure the classification accuracies of our method. Thirty people participated in the experiments, and six kinds of classification algorithms were used. As a result, the average of accuracies of the six classification algorithms was 97.87%.