The KanNon System - Vowel recognition using statistical pattern recognition

抄録

We have developed a real-time speech visualization system called “KanNon”[1, 2] which supports speech communication of hearing-impaired people. The KanNon system presents informations of the speech such as loudness, pitch, sound spectrogram, and characters by speech recognition system in real-time. In the present system, we are adapting word unit speech recongniton system using large scale dictionary. However the KanNon system requires quick and simple display of speech contents for smooth communication. For this purpose, we apply the methods based on neural network and statistical classfication for Japanese 5 vowels(/a/, /i/, /u/, /e/, /o/). Finally, we show results of phonemic recognition using real speech data.