On the Possibility of Measuring Learner’s Facial Expressions in On-Demand Learning Environments in Junior High School Students

In this research, we focus on the on-demand learning environment where learners study alone and examine the possibility of facial expression measurement as a method to grasp the state of learners. In the on-demand learning environment, it is considered that the learner’s emotions are less likely to be expressed in facial expressions than in general face-to-face learning or communication with others. Also, we don’t have sufficient information about what kind of emotions will represent. Therefore, in this paper, we construct a general facial expressions measurement system for measuring basic emotions, which appear commonly in all people, by image analysis and examine the possibility of grasping learners’ states from the facial expression analysis results. We extracted scenes in which the emotional states of the participants were assumed to be different and compared the results of emotional output in each scene. We adopted correct and wrong answers to the quiz task as a criterion for the differences in participants’ emotional states. The analysis results showed that the types of emotional output in scenes with different participants’ emotional states were similar. Still, there were differences in the output patterns of emotions. Based on these results, it is expected that a facial expression measurement system specialized for on-demand learning environments can be constructed by adjusting the facial expressions and emotions that appear in learning situations.