| 摘要 |
PURPOSE:To attain highly accurate voice input without checking the contents by editing recognized voice information as one record in accordance with a previously assigned procedure, and after checking the recognized result, inputting the record to a host computer. CONSTITUTION:Voice data from a microphone 1 are recognized by a voice recognition part 3 and the information of the recognized result is transmitted to a voice ansewering part 4 and an edition part 5 in any case, correct recognition or misrecognition. The answering part 4 combines the recognized result information with a previously set up voice answering comment, converts the combined result into a voice answering word and sends the voice answering word to an earphone 2. At that time, the operator can know the decision of correct recognition or misrecognition on the basis of the answering word. The recognized result information simultaneously sent to the edition part 5 is stored in a previously determined address. The edition part 5 stores successively sent recognized result information in respectively determined address to constitute a record. Finally, one stored record is sent to the host computer 8 on the basis of transmission instruction information to the computer 8 which is imputted as operator's voice.
|
