Speech and Language Technology for Disabled Persons
The Speech Training Aid Project is an investigation of the potential of speech recognition technology to aid and possibly screen young schoolchildren with speech problems. The system is based on a real-time re configurable continuous speech recognition system and uses sub-word models of speech to permit use of a large vocabulary. In use the child is prompted to speak words while the system monitors the child's speech The system then forms a report based on the number of acceptably pronounced words. At present the system utilises a vocabulary in excess of 1000 words, although the use of sub-word techniques gives the system the potential to accept any word for which a phonetic spelling is available.
Bibliographic reference. Series, R. W. (1993): "A speech training aid", In SLTDP-93, 173-176.