A voice-recognition system that enables conversion via subtitles has been developed for deaf participants who will attend an international conference of disabled people in Sapporo in October, according to the system's developers.
The system, to be used at the four-day Sixth World Assembly of the Disabled Peoples' International beginning Oct. 15, got rave reviews last year during a test in Sapporo, the developers said.
Jointly developed by Toru Ifukube, a Hokkaido University professor of welfare engineering, and systems development firm B.U.G. Inc., based in Sapporo, the system uses voice-recognition software developed by the Tokyo Research Laboratory of International Business Machines Corp.
Announcers in a separate room will input the remarks made at the assembly into the system, which will then instantaneously render them into Japanese or English subtitles on monitors showing the speaker's face.
The developers said there are still some errors in the Japanese translation where kanji are concerned but that the conversion to English subtitles is about 95 percent accurate.
During the test run, some were concerned about the faulty conversions, but said the system nonetheless helped them with note-taking. Even those who could hear said the subtitles enriched their understanding of what was said.
With further improvements in accuracy, it is hoped the system will allow disabled people to reduce their reliance on communications systems such as sign language, the developers said, adding that they are also hoping to adapt it to French and Spanish.