Attentional behavior model for a virtual character

Story-based VR systems are important for many applications, such as entertainment and education. To increase the reality of CG characters, character's attentional behaviors are important to guess what the character is interested in. In this paper, we propose a composite body-head-eye motion model for attentional behaviors. Attentional direction is divided into body, head and eye directions by a sharing mechanism. The head movement also generates the reflective eye movement. We apply the composite body-head-eye model to conversational scenes containing several people, and show that realistic and natural behaviors can be produced by using our model.