| 摘要 |
<p>PROBLEM TO BE SOLVED: To provide a speech system capable of accurately discriminating a speech from a suspicious person, and a speech method.SOLUTION: If a specific word is included in voice data in a speech, in a speech system 1A, it is determined whether voiceprint data of a speaking person is matched with voiceprint data of a preset speech permitted person and if not matched, risk of the speech is reported. In the speech system 1A, voiceprint data of speech permitted persons are registered as a white list in a voiceprint database 208, such that it is not necessary to register a number of unspecific voiceprint data in the voiceprint database 208 and that a speech from a third person who is not a suspicious person can also be prevented from being erroneously determined as a suspicious person.</p> |
