Factors Affect the Performance of Emotion Detection in a Classroom Monitoring System

抄録

Emotions play a key role in the way humans learn. However, understanding emotions in the context of online learning is complex, as many factors affect them. MOEMO (Motion and Emotion) is a classroom monitoring system that monitors online learners’ affective states and generates feedback to the teachers about the learners’ engagement and concentration level based on emotional data. The system captures data from a laptop or a desktop PC’s built-in camera. When capturing emotional data using the camera function, a few factors affected the performance of emotion detection. In this paper, we highlighted six factors that affected the result of emotion detection in the MOEMO system. Although these factors are common and, to some extent, inevitable, we aim to bring those to the learning analytics community.