A multi-purpose dialogue management system employing visual anthropomorphous agent

We have constructed a visual anthropomorphous agent system, called the visual software agent (VSA). The VSA has a texture-mapped realistic human facial image and moves his/her face with facial expression in real-time. Also, it can recognize the speech dialogue of a user and can respond in synthesized voice sentence by synchronizing his/her mouth movement. This type of multi-modal dialogue system may become a very important technology for advanced human interfaces in the near future. However, because our earlier system was designed to fit a specific application, it had been difficult to apply it to other applications. In order to solve this problem, we have constructed a multi-purpose dialogue system employing the VSA.