| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a voice recognition system that prevents equipment from remotely being operated by spoofing. <P>SOLUTION: The voice recognition system comprises an authentication data table 13 in which voiceprint information on respective users is registered as data for authentication, an acoustic analysis unit 11 which extracts voiceprint information and a feature quantity from a voice input from outside, an authentication processing unit 12 which retrieves whether the voiceprint information is registered in the authentication data table 13 and discriminates that authentication is completed when the voiceprint information is registered, a recognition decoder unit 14 which extracts a message from a prepared dictionary 15 by using the feature quantity as a key when it is discriminated to have been authenticated by the authentication processing unit 12, a command table 17 in which commands are registered according to respective messages in advance, and a message processing unit 16 which retrieves a command from the command table 17 by using the message extracted by the recognition decoder unit 14 as a key and operates a command 1 of the equipment with the command. <P>COPYRIGHT: (C)2007,JPO&INPIT</p> |
