| 摘要 |
PROBLEM TO BE SOLVED: To provide a speech recognition correcting system capable of efficiently correcting falsely recognized characters. SOLUTION: A speech recognition part 12 performs speech recognition processing to input speech by retrieving it from a speech recognition dictionary 14, and determines a character string data as the recognition result. When the character string data are erroneous, a user utters one character or (n) characters of corrective speech to a microphone 10, and a corrective speech recognition part 16 determines a corrective data corresponding to the corrective speech by speech recognition processing. A character replacement part 20 sequentially replaces a part of the character string data with the corrective data. A recognition result judging part 26 notifies the user of the character string data which have partly been replaced with the corrective data, and judges the character string data as a proper correction result according to a selection instruction by the user. |
