The REdRESS project

Recognition and Reconstruction of Speech following Laryngectomy

The recognition and reconstruction of speech based on audible signals is well developed but there are instances where audible signals are not available, such as following larynjectomy . We are developing a computer based system capable of detecting movement in the vocal apparatus which would allow the reconstruction of speech without acoustic signals. This has the potential benefit of providing improved speech quality and being less debilitating for patients than existing methods of restoration.

 This research builds upon expertise in Hull where it has been shown that indirect, non-acoustic measurement of speech activity is possible and work in Sheffield where speech recognition and synthesis has been used to construct intelligible speech from very limited speech signals, as in severe dysarthria, using a PDA.

• Detail
• Project Members
• Publications

The REdRESS project is funded by The Henry Smith Charity and Action Medical Research.