HITACHI : Intelligent User Interface Based on Multimodal Dialog Control for Audio-visual Systems

March 2006

AUTHORS

Hiroshi Shinjo	Intelligent Media Systems Research Dept., Central Research Lab., Hitachi, Ltd.
Ui Yamaguchi	Intelligent Media Systems Research Dept., Central Research Lab., Hitachi, Ltd.
Akio Amano	Intelligent Media Systems Research Dept., Central Research Lab., Hitachi, Ltd.
Konagi Uchibe	Intelligent Media Systems Research Dept., Central Research Lab., Hitachi, Ltd.
Atsushi Ishibashi	Home Solution Design Gr., Design Div., Hitachi, Ltd.
Hideki Kuwamoto	R&D Strategy Planning Dept., Business Planning Office, Ubiquitous Platform Systems, Hitachi, Ltd.

OVERVIEW

As the functionality of AV equipment improves and the number of TV broadcast channels increases, the era in which we "can't pick and can't watch" our preferred programs is approaching. In addressing this problem, Hitachi has developed an "intelligent user interface"—which uses "multimodal dialog control" to help users (i.e. viewers) operate home AV equipment by means of spoken dialog. Integrating speech information processing, image information processing, and a TV program recommender based on viewing-history analysis, this multimodal dialog control can partake in dialog with the viewer. Without the need for keyword entry, the TV program recommender automatically analyzes the viewer's program preferences from their past viewing history, and then selects, recommends, and records certain programs according to the analysis results. Utilizing voice information processing, this intelligent user interface can operate AV devices via spoken dialog very close to everyday conversation. It can also distinguish different viewers by facial image recognition and automatically change the offered services (such as the TV program recommender) to match the preferences of whoever is in front of the TV.

TO READ THIS ARTICLE

See PDF file for details(PDFformat, r2006_mar_002.pdf, 126KB)

RELEVANT SITES

	News Release Digest (PDFformat)
	News Release (Japanese Only)
	Central Research Laboratory
	The Hitachi Hyoron (Japanese Only)

KEYWORDS

multi-modal human interface, voice synthesis, voice recognition, face recognition, program recommendation

PDF files are in Adobe's Portable Document Format. To view them you need Adobe(R) Reader(TM) by Adobe Systems Incorporated.