Embodied Conversational Agents for Presenting Intellectual Multimedia Contents

摘要

This paper presents an embodied conversational agent (ECA) that presents multimedia contents. The system takes plain text as input, and automatically generates a presentation featured with an animated agent. It selects and generates appropriate gestures and facial expressions for a humanoid agent according to linguistic information in the text. As a component of the ECA system, we also present an agent animation system, RISA, which can draw animations of natural human behaviors on web-based applications.