Input support system via voice for a physically handicapped person with an utterance obstacle

摘要

We have designed and implemented an input support system via voice. Our system enables a physically handicapped person who has also trouble in utterance to operate a personal computer. Commercial speech recognition products are not applicable as it is, because their dictionaries include too many words and his ambiguous utterance often causes mis-recognition. We have investigated several sets of words whose recognition is possible with his cooperation. Then we have defined input modes and built categorized dictionaries so as to specify variety of words from limited words. Through an input experiment, we have confirmed that the physically handicapped person with an utterance obstacle is able to input by using our system.