| 摘要 |
PROBLEM TO BE SOLVED: To provide an utterance difference speech recognition system capable of recognizing speech with a high recognition rate, even in speech included utterance difference, which is deviated from standard utterance, due to regional difference etc. SOLUTION: A recognition computer 2 of the recognition system 1 includes: a speech information acquiring means 9 for acquiring speech information 8; a dictionary group storage means 10 for storing a standard dictionary SD constructed so as to recognize the standard utterance, and an extension dictionary ED constructed so as to recognize non-standard utterance including utterance difference; a standard collation means 11 which collates a vocabulary included in speech, and recognizes a vocabulary which matches to a vocabulary in assumption, by using the speech information 8 and the standard dictionary SD; an extension collating means 12 which collates a vocabulary which does not match the assumed vocabulary by the standard collation means 11 by using the extension dictionary ED, and recognizes a vocabulary which matches to a vocabulary outside assumption; and a vocabulary output means 13 for outputting the vocabulary recognized by the standard collation means 11 and the extension collating means 12. COPYRIGHT: (C)2010,JPO&INPIT |
