Time-Sequential Emotion Diagnosis Based on the Eye Movement of the User in Support of Web-Based e-Learning

Abstract

Learning processes are becoming increasingly mediated by computers and the internet. E-learning, especially Web-based learning, is penetrating into various fields. An ordinary e-Learning system usually evaluates learners’ states from their inputs to the pre-programmed questions. The time-sequentially changing emotions of learners are important for the management of the learning environment, just like face to-face classes. The present paper describes an emotion diagnosis system that can judge the emotions of an e-Learning user based on his/her eye movement speed and fixation duration time. The criteria for classifying four pairs of semantically different emotions (eight emotions in total) were established through a time-sequential subjective evaluation of the emotions of the subject and a time-sequential analysis of the eye movements of the subject. The achieved coincidence ratios between the discriminated emotions based on the criteria of emotion diagnosis and the time-sequential subjective evaluation for “tired” and “concentrating” were 77% and 60%, respectively.