摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a device and a method for speaker recognition which are robust to false names, background noise, etc. <P>SOLUTION: The speaker recognizing device is equipped with an input means (4) of inputting a voiceless utterance of a speaker and speaker recognizing means (1, 61) of recognizing the speaker according to the inputted voiceless utterance. In a preferable embodiment, the speaker recognizing means perform 1st speaker authentication based upon the voiceless utterance inputted from the input means and 2nd speaker authentication using the inputted voiceless utterance as password input. <P>COPYRIGHT: (C)2005,JPO&NCIPI</p> |