3D Animation using Visual Script Language

In animating virtual characters, it is necessary to provide an interface that allows users specify the characters’ motion with high-level concepts as well as low-level details. And it will be convenient to user if the system provides an animation agent that processes various tasks for animation. In this paper, we suggest an agent (HUman-MOtion Agent: HUMOA) that controls behaviors and actions of virtual characters automatically. The agent is structured with multi-level and is operated according to semantics of scenario that is composed of events and visual scripts. Visual script is the gesture with mouse that describes characters’ motion visually. The agent of multi-level structure has internal knowledge that is expressed as concepts and relations. So the agent can perform tasks of each level of abstraction automatically. A user can interact with virtual characters directly and intuitively using visual script language. And the agent of our system makes it easier to design animation scenario and to perform tasks for animation compared with traditional direct manipulation or motion control by