Nagoya Institute of Technology, Center for Speech Technology Research

Nagoya Institute of Technology, Center for Speech Technology Research
The MMDAgent Speech Synthesis Kit for Real-time Conversation with a CGI Character.
The Nagoya Institute of Technology, Center for Speech Technology Research presented a demonstration introducing their prototype MMDAgent software and speech interaction system. Via the MMDAgent software, visitors to the booth were able to engage in real-time conversation with a CGI character displayed on the screen. The exhibit also featured demonstrations of dialogue with the virtual diva Hatsune Miku.
Freeware that brings together voice recognition and synthesis technologies MMDAgent software is based on four different technological elements - the HTS voice synthesis kit, a Julius voice-recognition engine developed and released by NIT's Center for Speech Technology Research, a 3D rendering module, and a dialog module. NIT also plans to release MMDAgent as freeware once it has passed the prototype stage. The speech interaction control section can be configured to suit a wide range of actions (conversations, motions, etc.) in respond to both internal and external stimuli such as speech input. What's more, as the descriptions for various dialogue scenarios are written in a script format, general users who do not have specialist knowledge can freely configure the software. OpenGL 3D rendering functions ensure high performance to enable imaging with abundant toon rendering (cel-shaded animation) and shadow mapping in addition to providing life-like expressions via the physical engine. Because all model data is in the open-source format, everything from the 3D character models and motions through to speech and dialogue scenarios can also be freely customized. Therefore, when this freeware released, it is sure to become a highly attractive tool for all creators from general users to specialized professionals.
Strong interest from the tourism industry and an important tool for establishing venture enterprisesApplying specialized research into speech technologies, NIT's Center for Speech Technology Research's reputable high-precision tools have been developed based on its vast amount of accumulated knowledge and expertise. In addition to generating significant interest internationally, NIT has received many inquiries from customer service businesses and the tourism industry throughout Japan and notably, its technologies are used in mobile device applications in China as well as in car navigation systems. The venture enterprise, "Techno Speech", was established within NIT to enable it to better delineate between its objectives as a research institution advancing technology and its involvement in corporate economic activities. This will enable these highly flexible development tools to be provided to a wider range of users and at the same time, allow NIT to focus on supporting the development of even higher precision software to meet the needs of society in general. Using a vertically mounted screen and microphone, NIT's demonstration at CEATEC Japan 2010 featured a life-sized character answering questions in real time with the virtual diva, Hatsune Miku, also making an appearance.

Booth number: 2B07