| 摘要 |
PROBLEM TO BE SOLVED: To provide a system capable of performing voice recognition by improving voice section detection accuracy, even in an environment in which noise level fluctuates rapidly. SOLUTION: This system has a first filtering means (2) for reducing noise component, with attention given thereto from an input signal (1), a second filtering means (3) for extracting the noise component with attention given thereto from the input signal, and a voice section detecting means (7) for detecting the voice section of the input signal, on the basis of outputs of the both filtering means. The voice section detecting means detects the voice section by comparing first signal information obtained, on the basis of the output of the first filtering means with a threshold, and a recognizing means (15) performs voice recognition in the voice section. The voice section detecting means can change a threshold for voice section detection by making the threshold follow change of a noise environment because the voice section detecting means changes the threshold according to the fluctuation of second signal information obtained, on the basis of the output of the second filtering means. |
