摘要 |
PROBLEM TO BE SOLVED: To provide a voice recognition apparatus capable of reducing delay of voice recognition processing by reading a word dictionary from an auxiliary storage device in a device with a memory capacity restricted. SOLUTION: The voice recognition apparatus for recognizing a plurality of successive words included in voice, comprises: a sound model reading section 5 for reading a sound model 11; a dictionary managing section 9 for reading a necessary data from a dictionary data 12; and a recognition section 7 for sequentially recognizing the successive words by collating a word group expressed by the dictionary data 12 to the input voice by using the sound model 11. The dictionary data 11 include a word head dictionary for indicating a word head of the word, and a word end dictionary group in which the data for indicating a word end are recorded, and reads the word end dictionary and/or the word head dictionary, according to the word recognized by a recognition section 7. The recognition section 7 collates a following voice to the word head of the word included in the word head dictionary, while a dictionary managing section 9 is reading the word end dictionary and/or the word head dictionary. COPYRIGHT: (C)2008,JPO&INPIT
|