Continuous Interaction with a Virtual Human

Attentive Speaking and Active Listening require that a Virtual Human be capable of simultaneous perception/interpretation and production of communicative behavior. A Virtual Human should be able to signal its attitude and attention while it is listening to its interaction partner, and be able to attend to its interaction partner while it is speaking – and modify its communicative behavior on-the-fly based on what it perceives from its partner. This report presents the results of a four week summer project that was part of eNTERFACE’10. The project resulted in progress on several aspects of continuous interaction such as scheduling and interrupting multimodal behavior, automatic classification of listener responses, generation of response eliciting behavior, and models for appropriate reactions to listener responses. A pilot user study was conducted with ten participants. In addition, the project yielded a number of deliverables that are released for public access.