摘要 |
PROBLEM TO BE SOLVED: To provide voice recognition result optimizer capable of optimizing a voice recognition result via a keyword list without modifying voice recognition engine with a small calculation cost.SOLUTION: The voice recognition result optimizer includes: a reception section that receives a voice recognition result from a voice recognition engine; a calculation section that calculates the similarity in a voice between a segment of the voice recognition result and a keyword in a keyword list; and a replacement section that, when the similarity of the voice is higher than a first threshold level, replaces the segment with the keyword.SELECTED DRAWING: Figure 1 |