[4F3-OS-25b] OS-25 (2)

[4F3-OS-25b-02] Human emotion recognition vs machine emotion recognition using spontaneous and posed expressions

〇Shushi Namba¹ (1. Hiroshima University)

Keywords:Facial expressions, Emotion

How far does machine (i.e., AI) achieve emotion recognition from facial expressions compared to human performances? This study aimed to compare human performance and machine analysis (FACET) for emotion recognition of facial expressions. Previous studies mainly used only posed expression stimuli, thus the current study tried to include spontaneous expressions in addition to posed ones. The results showed that the mean values of emotion recognition accuracy did not have differences between human performance and machine analysis. However, the Bayesian modeling approach identified the specific interpretation pattern by human compared to machine analysis. By considering two patterns of emotion recognition among human and machine, this study discussed about what will contribute the affective computing fields.