Abstract
For many years, Ekman’s research on facial expressions has been the basis of research on facial expression recognition or estimation. However, in recent years, it was reported that Japanese people do not follow Ekman’s theory, and much remains to be clarified about the relationship between emotion and facial expression. In this study, we focused on the natural expression of eye region, which is unintended expressions of emotions, and analyzed the characteristics of it in comparison with facial features expressed in spontaneous emotional arousal situations. As a result, we revealed that the size of the eyes differed for each emotion and that the expression of the eye region varies greatly among individuals. In addition, we suggested that the mouth may be useful for pleasantness estimation by machine learning.
