Examinee Authentication Using Multiple Types of Touch Gestures for Proxy Detection in Online Examinations

Abstract

This study proposes a new method for detecting and preventing proxy test-taking in online examinations using a tablet device that authenticates the identity of examinees based on multiple types of touch gestures. We proved that the support vector machine has the best performance among classifiers for authentication when various touch gestures are used and various positions on the screen are touched. The results also showed that if support vector machine is used as the classifier, a handshape image taken from above or from the side can be used for identity authentication. In the future, we will improve the experimental environment to make it more similar to an actual online examination and investigate the possibility of performing authentication using touch gestures at arbitrary points on the touch screen of a tablet device.