Acted emotional expressions of game-playing children: investigating the influence of emotion intensity on recognition rates

While the intensity of emotions is likely of great importance to automatic emotion recognition systems, it is not an ordinary feature in emotion databases. This paper presents a database of children acting out six basic emotions, in which the intensity of said emotions was manipulated. A judgment task showed emotions were better recognized than chance could predict, while differences in intensity were perceived in two manipulated conditions. In addition, we establish a correlation between intensity and recognition rates for this database. Finally, possibilities for future research are discussed.