| 摘要 |
PROBLEM TO BE SOLVED: To provide a speech recognition device and a speech recognition method that can correctly recognize a spoken speech by correcting a wrong recognition error of a keyword even if the keyword is incorrectly recognized. SOLUTION: A previous-stage recognition unit 102 regards a set of keywords included in a plurality of vocabularies to be recognized in common as an expected word and recognizes a keyword included in a spoken speech input through a speech input unit 101. A previous-stage recognition result presenting unit 105 presents the recognition result of the keyword to a user, and a correction receiving unit 106 receives a correction of the presented recognition result to correct the recognition result of the keyword. A dictionary switching unit 104 structures an expected word dictionary for recognizing the spoken speech while taking the correction result into consideration, and a post-stage recognizing unit 108 recognizes the spoken speech using the expected word dictionary. COPYRIGHT: (C)2008,JPO&INPIT
|
