| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a conversation system capable of accurately determining what a name voiced by a user during a conversation indicates and of returning an appropriate answer. <P>SOLUTION: The user's voice inputted through a voice input part 12 is imparted to a CPU 11 via an A/D converting part 13. The CPU 11 recognizes the user's voice, thereby detecting the name of a specific person therefrom. The CPU 11 identifies the person indicated by the name based on the similarity between the name detected from the user's voice and the name of each person stored in a nonvolatile memory 17 and returns a remark about that person as an answer to the user. <P>COPYRIGHT: (C)2005,JPO&NCIPI |
