摘要 |
PROBLEM TO BE SOLVED: To provide a technology of improving the accuracy of speaker recognition.SOLUTION: The voice processing device includes: acoustic variety calculating means for calculating, on the basis of a voice signal showing a voice, acoustic variety indicating the degree of variation related to the type of sound included in the voice signal; and feature correcting means for correcting the recognition feature amount calculated for recognizing specific attribute information from the voice signal using the acoustic variety of the voice signal.SELECTED DRAWING: Figure 1 |