Automatic Detection of Mind Wandering Using Mobile Device In the Scenario of Online and Mobile Education

Mind wandering is a state transition from task-related thinking to task-unrelated thinking and it has possible negative effects on learners’ performance in online education. The detection of mind wandering can improve learners’ performance by signaling, intervening and restoring attention to the process of learning. Traditional methods of detection of mind wandering involve specific devices such as eye gaze or physiologic sensors, tracking movements of eye ball or detecting electro-physiological activities of neurons or skin. This paper presents a new detection method using the camera of a mobile device to detect mind wandering when learners are tasked to watch a teaching video. Through the constant image capturing of the learner’s face, his or her eye features can be captured and used to predict the learner's state of mind wandering if any, using methods of computer imagery and supervised machine learning. Through our experiment, we have attained an accuracy of 53%, which exceeded expectations, considering the hardware platform. The performance of the detection system based on the MCPP mechanism outperforms the detection system based on the traditional mechanism.