| 摘要 |
PURPOSE: To improve a recognition rate and to enable the interaction meeting the change in condition and environment by executing the recognition taking the change in condition and environment into consideration. CONSTITUTION: This interactive device has a timer means 3 which generates, for example, time data as fluctuation data and a coeff. setting means 4 which is inputted with the time data from this timer means 3 and generates the weighting coeffs. changing with a change in the time data corresponding to the contents of the respective speeches to be registered and recognized. The device has a speech recognition means 5 to which the speed data patterns of the input speeches outputted from a voice analyzing means 2 are inputted, obtains the respective weighting coeffs. at the point of this time for the speeches to be registered and recognized from the coeff. setting means 4, calculates the final recognition data taking the respective weighting coeffs. into consideration and executes the recognition of the input voices from the calculated recognition data. The device has a speech synthesis means 6 which outputs the speech synthesis data in accordance with the recognition data taking the coeffs. from the speech recognition means 5 into consideration and a speed output means 7 which emits the output from this speech synthesis means 6 to the outside. |
