摘要 |
PROBLEM TO BE SOLVED: To provide a device and a method for speech recognition that can present an optimum instruction word to a user to surely instruct an operation. SOLUTION: Frequencies of misrecognition of respective words in a word list are counted from a result of speech recognition of a user's utterance including utterance not to be recognized to exclude a word, selected according to a first standard based upon the frequencies of misrecognition (404), from an object to be recognized, and frequencies of correct recognition of the respective words in the word list are counted from a result of speech recognition of user's utterance in the object to be recognized to exclude a word, selected according to a second standard based upon the frequencies of correct recognition, from the object to be recognized (405). COPYRIGHT: (C)2007,JPO&INPIT
|