08-03-2017, 12:38 PM
Voice Communication support output voice input (VIVOCA)
The goal of the project is to develop a portable communication aid for people with disordered or unintelligible language, initially focusing on people with moderate to severe dysarthria (people who have difficulty controlling and coordinating the muscles used in speech). The device will effectively translate speech from a disartric person into clear voice output (synthesized or recorded). This will be useful for dysarthric people in social situations in which they interact with non-family communication partners, such as at work, shopping, hospital, telephone, etc. There is ample evidence to demonstrate that such a device is necessary and would be beneficial. The intention is to provide the user with a device that acts as a human interpreter, recognizing disordered speech and producing an intelligible equivalent.
The work is based on the STARDUST (NEAT A230) project, which successfully developed a way for discordant people to control speech technology, but it requires many innovative steps to address this particularly challenging problem, including the development of :
• Techniques to optimize variable voice recognition in noisy environments
• Interpretation and translation of speech in a natural and acceptable way
• A means to enable help to be portable (including the body) and to help face-to-face natural communication
• Adaptive techniques as speech and user needs change over time, even for people with progressive neurological disease (eg, MND)