COMPUTER SYSTEM FOR SPEECH RECOGNITION

摘要

PURPOSE: To provide a computer system for recognizing voice most simply and speedily executing voice recognizing processing with respect to an unknown word. CONSTITUTION: In response to that the computer system comes across with a new word which is not included in a word table A in the middle of recognizing an unknown pronunciation text, the phonemic basic form of this new word is prepared synthetically by the support of a triphone table, the new word is decomposed to triphones to read phonemic sequence related with an individual triphone, and the phonemic sequence is characteristically connected according to the phonemic basic form of the new word.