ABSTRACT

The present paper describes an interactive multimodal motion learning system for traditional arts. This system uses several modalities, such as synthesized speech, video, and agent actions, to instruct a user. Input modality uses speech modality to control the system considering a hands-busy situation in motion learning. This system is developed using the proposed multimodal interaction system framework, which realizes easy construction of learning contents from high-level data modeling. As an example to illustrate the use of this framework, we apply the multimodal learning system to the Japanese tea ceremony.