相手顔画像を合成した身体的引き込み観客キャラクタを用いた実映像対話システム

抄録

An embodied video communication system, called 'E-VChat', was developed for enhancing remote communication using a video image. This system is constructed by the partner's video image using interactive CG characters including user's own avatar and virtual audience characters. We have confirmed the effectiveness of the system by an experiment under free conversation. In this study, we focus on consensus building communication using a video communication system with user's own avatars and virtual audiences. For consensus building from different opinions of speakers, we can expect the system to provide confidence to speakers about their own opinions using the interactive CG audiences, because speakers can experience the agreeable responses from virtual audiences. We have developed embodied communication systems including two users' avatars and speech-driven embodied entrainment objects using plant or shadow-figure models, and the effectiveness of this communication system which presented entrainment reaction was confirmed in the experiment under consensus building. Accordingly, we develop an embodied entrainment audience characters system with partner's face to relate audience characters with the partner. This system could bring about more active interaction by virtual audience even if the remote partner does not respond to user's speech. And we thus conduct two communication experiments using this system to confirm the effectiveness of the system for 12 pairs of 24 talkers.