摘要 |
PROBLEM TO BE SOLVED: To provide a speech recognition device excellent in recognition accuracy of speech instructions including a common words. SOLUTION: The speech recognition device which pre-stores two or more recognition dictionaries including a predetermined recognition vocabulary corresponding to operation states of two or more equipment to be controlled, which selects at least one recognition dictionary based on the operation states of the two or more equipment to be controlled, and which recognizes a speech instruction uttered by a user based on the selected recognition dictionary, judges (S50, 60) whether or not the common words included in at least two recognition dictionaries in common among the two or more recognition dictionaries is included in the speech instruction uttered by the user; narrows the recognition dictionaries down to one of them (S80) based on the information having stored the equipment operated last (last management information) by the user when the speech recognition device determines that the common word is included; and thus identifies what the speech instruction uttered by the user is, by performing speech recognition processing using the one recognition dictionary. COPYRIGHT: (C)2007,JPO&INPIT
|