An affect-expression space model of the face in a mascot-type robot

Abstract

Humans want robots which can react or express their emotions similarly as human beings do. However, there has not been any definite models to express emotion on the face of robots. The previously developed face of robots just has some target positions of the facial components which were defined in advance for expressing, and it can not show the continuous emotional expression while its emotion changes. In this paper, we will see the definition of an expressional space on the face of a mascot-type robot and its application to the formation of an affect-expression space model. The proposed model is simple but more efficient than other methods since it can be used to express various and continual emotional expressions just using a little facial information. The results will be shown in experiments using a mascot-type facial robot system.