| 摘要 |
PURPOSE:To evade a decrease in recognition rate. CONSTITUTION:This voice recognition device is provided with three ultra- directional imcroiphones 11-13, three speech recognition parts 51-53, peak holding circuits 91-93, a CPU 7 which controls them, etc. Each speech recognition part compares voice data obtained from a microphone with previously registered voice data to perform speech recognition. Each peak holding circuit holds the peak value of voice data and sends it to the CPU 7, which outputs a recognition result from the speech recognition part of a system having the highest peak value, so that the recognition rate is improved even in the presence of a loud noise. |
