| 摘要 |
PROBLEM TO BE SOLVED: To provide a multi-layered speech recognition apparatus and a multi-layered speech recognition method. SOLUTION: The multi-layered speech recognition apparatus is provided with a client 10 which has a function to determine whether a speech can independently be recognized for the speech features to be recognized or not and the speech is recognized in accordance with the decision result or to transmit the features and a first server 20 to an N-th (where N is equal to one or greater number of positive integers) server 24. The first server 20 decides to determine whether the speech can be independently recognized using the speech features transmitted from the client 10 or not and recognizes the speech in accordance with the decided result or transmits the speech features. An n-th (2≤n≤N) server, which is any one of the servers to the N-th server excluding the first server 20, decides to determine whether it is possible to independently recognized the speech for the speech features being transmitted from the first server 20 or an (n-1)-th server or not and recognizes the speech based on the discriminated result or has a feature to transmit the speech features. COPYRIGHT: (C)2006,JPO&NCIPI |
