| 摘要 |
PROBLEM TO BE SOLVED: To present confirmation information for confirmation results in a manner easy for a user to understand while allowing recognition of an optional utterance by voice recognition using subword grammar. SOLUTION: A voice recognition part 102 performs, to an input voice, voice recognition using keyword grammar based on a keyword for a state related to registration, and voice recognition using subword grammar based on a subword, and outputs recognition results of the recognition having a more satisfactory recognition score of the both. When the voice recognition part 102 outputs the recognition results of the voice recognition using the keyword grammar, a confirmation information generation part 106 generates confirmation information composed of a keyword string of the recognition results. At the same time, when the recognition part 102 outputs the recognition results of the voice recognition using the subword grammar, the generation part 106 generates confirmation information composed of attribute information associated with the state related to registration. COPYRIGHT: (C)2009,JPO&INPIT
|
