| 摘要 |
PROBLEM TO BE SOLVED: To provide a voice recognition device, etc. capable of easily and exactly correcting the erroneously recognized utterance content when the uttered content is erroneously recognized.SOLUTION: A voice recognition device accepts, after a voice recognition result (voice character strings) of the utterance content of a driver, etc. is displayed, input of a correction character string which is a normal voice recognition result of an erroneously recognized part (erroneously recognized character string) in the voice character strings from the driver via a touch pad attached to a steering (S230). Then, character strings with the same number of characters as that of the correction character string are extracted from the voice character strings, a character string with the most characters matched to those of the correction character string is detected as the erroneously recognized character string among the character strings (S240), and the voice character strings are corrected by replacing the erroneously recognized character string with the correction character string (S265). |
