Human Interface employing Intelligent Anthropomorphic Agent with Realistic Figure

人机界面技术在先进的信息化社会中变得越来越重要。作为一种新型的人机界面，我们开发了一种基于实时图像合成、实时运动图像识别、语音识别/合成和知识库技术融合的拟人化代理系统VSA（visual software agent）。我们的VSA具有通过实时纹理映射到3D模型生成的逼真面部图像，语音通信能力，响应人类用户视觉识别结果的交互能力，以及对Internet/WWW的访问功能。特别是，我们已经成功地将我们的VSA与Internet中典型的WWW （World Wide Wed）浏览器Mosaic连接起来。因此，我们的VSA系统具有很大的优势，它可以利用互联网上庞大的多媒体信息源，并且通过VSA接口传输的新信息可以用广泛使用的标准化HTML（超文本构成语言）格式编写。为人类日常生活提供一种接近面对面交流的新型多模态WWW浏览器成为可能。

Human interface technologies are becoming important increasingly in the advanced information-oriented society. As a new-type human interface, we have developed an anthropomorhipc agent system called VSA(visual software agent), which is based on the fusion of realtime image synthesis, realtime moving image recognition, speech recognition/synthesis and knowledge base technologies. Our VSA has a realistic facial image generated by a realtime texture-mapping onto a 3D model, speech communication capability, interactive capability responsive to the visual recognition result of a human user, and an access function to Internet/WWW.In particular, we have successfully connected our VSA with Mosaic, a typical browser of WWW (World Wide Wed) in the Internet. As a result, our VSA system becomes to have a big advantage that it can use vast multimedia information sources in the Internet and new information to be transmitted through the VSA interface can be written in a wide-spreaded standalized HTML (hyper text makeup language) format. Also, it becomes possible to provide a new-type WWW browser with new multi-modalities close to face-to-face communication in human daily life.