Video conference smart room: an information fusion system based on distributed sensors
Wan Khairunizam, Atsushi Todo, Hideyuki Sawada, Olivier Passalacqua, Éric Benoît, Marc-Philippe Huget, Patrice Moreaux · 2008
Abstract — The needs for cross domain technologies called mecatronics have increased in the recent years and examples of mechanical tools directed by computers are now widely available. This paper gives an example of information fusion in the context of a video conference room and exposes two axes of the research. First it shows how to fuse information provided by several sources to locate a speaker. To do this, the system fuses data produced by video cameras and their associated image processing algorithm, with information resulting from signal processing algorithms applied on several micro-phones. Second, this article describes the distributed information fusion system (DIFS) used and the algorithm which decides where the speaker is located in order to allow focus on him. The whole application is managed by a new control system specifically developed for DIFSs. Some key points of the theoretical model on which the control is based are also given. This project has been realized thanks to an international cooperation between the Kagawa University in Japan and the University of Savoie in France. KEYWORDS distributed system, image and signal