| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a voice recognition device that makes voice recognition processing by performing the voice recognition processing in parallel. <P>SOLUTION: The voice recognition device includes a voice division part, a voice storage part, a division voice distribution part, a voice recognition processing group, and a voice recognition result integration part. The voice division part inputs a voice signal, makes a voice/non-voice determination on the voice signal, and divides a voice section when the section length of a non-voice section is equal to or larger than a non-voice division section length threshold to output division voices obtained by dividing the voice signal into utterance section units. Then the division voice distribution part distributes utterance sections stored in a division voice storage part to a voice recognition processing group comprising a plurality of voice recognition parts. The voice recognition result integration part combines a plurality of voice recognition results that the voice recognition processing group output, and outputs a voice recognition result for the voice signal. <P>COPYRIGHT: (C)2013,JPO&INPIT</p> |
