| 摘要 |
PURPOSE:To rapidly and exactly execute the recognition processing of voices by executing such a preprocessing as noise removal of an input voice signal with CPU groups of lower priority hierarchy and sharing the execution of the pattern matching processing for the purpose of voice recognition with CPU groups of upper priority hierarchy. CONSTITUTION:A voice input section 10 consists of a suitable microphone 10A and a microphone amplifier 10B. This device includes characteristic extracting sections 11-1 to 11-4 which extract the characteristic parts in the waveform of the input voice signal and a data processing section consisting of plural pieces of the CPUs 15-1 to 15-4, 16-1 to 16-2, 17 connected by having the hierarchy structure. Such preprocessing as noise removal relating to the input voice signal is executed by the CPU groups 15-1 to 15-4 constituting the lower priority hierarchy and the execution of the pattern matching processing for the purpose of the voice recognition is shared with the CPU groups 16-1, 16-2, 17 constituting the upper priority hierarchy. The voice recognition processing is rapidly and exactly executed in this way. |
