| 摘要 |
PROBLEM TO BE SOLVED: To actualize speech recognition of high certainty by effectively decreasing the number of vocabularies as objects of speech recognition. SOLUTION: In steps S201 to S203, a speech feature quantity is extracted from an inputted speech signal and the speech signal is divided into partial speech signals according to the speech feature quantity. In a step S207, speech recognition is carried out by the partial speech signals obtained in the step S203, and object vocabularies, i.e., recognition dictionaries to be used for the speech recognition are changed for every partial speech signal (step S213).
|
