(image: Pixabay)
SEOUL, Oct. 7 (Korea Bizwire) — South Korean researchers have developed a silent speech recognition system to help hearing-impaired people communicate, enabling users to understand what others say simply by recognizing the changes in their lip shape.
A research team from Yonsei University said Thursday that it had developed a silent speech recognition system combining a monocrystal silicon-based high-performance wearable strain gauge with a deep learning-based strain-word shift algorithm.
The system can identify words at an accuracy rate of up to 90 percent without actual voice by effectively analyzing facial movements.
It is expected to enable people with hearing and speech impairments to communicate effectively simply through their lip shape, while being useful even in military and police operations and security systems.
“This is a new-concept platform that enables handicapped people to communicate by learning the motion of lip shape without using sign language,” the research team said.
J. S. Shin (js_shin@koreabizwire.com)
