| 摘要 |
PROBLEM TO BE SOLVED: To realize the improvement in operability and responsiveness, even when the input of a narrowing-down condition which a user does not know is required by a system when inputting a narrowing-down condition hierarchically to performing point designation, by developing a structure which enables proceeding to a subsequent step for performing the point designation. SOLUTION: Reference voice information is extracted from among a plurality of hierarchized speech recognition dictionaries. Reference voice information showing hierarchical jumping is prepared in a prescribed speech recognition dictionary about speech recognition for recognizing speech by comparing the extracted reference voice information with inputted speech. When an input corresponding to the reference voice information showing hierarchical jumping has been recognized, a part of speech recognition dictionaries belonging to the low order hierarchy of the reference voice information subjected to the comparison are extracted to perform the speech recognition. |
