Development of an interactive robot with emotional expressions and face detection

The purpose of this paper is to develop an interactive robot, named Golden Horn, which can display emotional expressions and detect human faces. With the 12 degrees of freedom, Golden Horn can generate rich emotional expressions. In addition, a face-detection algorithm, based on the knowledge-based method, is proposed. It adopts the geometric characteristics of skin and hair colors to detect the human faces. We have conducted several experiments to test the developed robot. Experiments showed that emotional expressions generated by Golden Horn can be identified well. In addition, the accuracy of the single-face detection is higher than 92% with the frame-rate of 10 fps. The developed robot shows great potential to be applied for social interaction with the humans.