| 摘要 |
PROBLEM TO BE SOLVED: To provide a speech processing method and system capable of surely urging reutterance even under surrounding large noise. SOLUTION: The speech processing system is equipped with a speech input section 1 for inputting human speech, a speech recognition section 2 for recognizing the inputted speech, a recognition result decision section 3 for deciding whether the recognized result can be correctly recognized or cannot be recognized or that the recognized result is error and a recognition result presentation section 4 which presents the recognized result to the human. The system is equipped with an acoustic data setting section 5 for forming or selecting the acoustic data easiest to catch under the surrounding noise environment and an acoustic data storage section 6 for storing the formed or selected acoustic data. The acoustic data set in the section 5 is outputted as sound to the section 4 when the result of the speech recognition is an erroneous recognition or a recognition failure in the section 3. COPYRIGHT: (C)2004,JPO&NCIPI
|
