| 摘要 |
PROBLEM TO BE SOLVED: To acquire voice print data for voice print authentication in real time from received voice in communication without drawing attention to voice of a conversation partner, automatically update a determination threshold set individually for determining from the voice print whether a person is the person who owns the voice print, and notify a user of the determination result.SOLUTION: Received voice is acquired, the telephone number of a speech side is acquired, and voice print data is extracted from the acquired received voice. Further, the acquired telephone number and the extracted voice print data are correlated to each other and are stored if the acquired telephone number is in a phone book. A score is calculated indicating the similarity between registered voice print data and the voice print data extracted from the acquired received voice. Further, whether the voice print data extracted from the acquired received voice is correct or not is determined by comparing the calculated score and a threshold individually set in advance, and the determination result is notified if the determination result is not correct. The individual threshold is updated every time the correctness is determined.SELECTED DRAWING: Figure 2 |
