| 摘要 |
<p>PURPOSE: To detect a word by determining a new sequence number in each comparison between a new voice value and the final reference value of at least one word and searching a 3rd list entry group related to the sequence number of the word stored in a 2nd list in each comparison. CONSTITUTION: The compared result of a word obtained by practically comparing a sound voice value supplied through an input terminal. 12 with a reference value is supplied to a section 2 through a pass 8, respective words are stored and the always completely recognized compared results of these words are used together with grammar rules. A succeeding new word is determined based upon the grammar rules and supplied to a section 1 through a pass 11 and a reference value to be compared with a succeeding voice value to be supplied through the input terminal 12 is determined or plural reference values are prepared. Thus voice recognizing processing is always alternately executed between the sections 1, 2 until the final voice value is compared. Consequently a word can be detected.</p> |
