| 摘要 |
PURPOSE: A voice recognition device and a method thereof are provided to offer a high performance voice recognition service without increasement of a terminal manufacturing cost, and to increase a service channel per a server by decreasing the calculation quantity of a server through removing the degradation of voice signal and extracting a characteristic in a terminal. CONSTITUTION: A voice recognition device includes a characteristic extracting part(2), a voice signal processing part(5), the first switch(SW1), a CPU(6), the second switch(SW2), and a server(200). The characteristic extracting part(2) receives a voice of a user according to a mode control signal of the CPU(6) in a recognition mode, and extracts a characteristic of the voice. The voice signal processing part(5) processes the voice according to a mode control signal in a calling mode. The first switch(SW1) is switched to the characteristic extracting part(2) or the voice signal processing part(5) according to the mode control signal. The CPU(6) outputs the mode control signal according to a function selection of the user, and transforms an output signal of the voice signal processing part(5) to transmission data, or outputs an output signal of the characteristic extracting part(2) as it is. The second switch(SW2) is switched in order that the output signal of the characteristic extracting part(2) or the voice signal processing part(5) is applied to the CPU(6). The server(200) includes a voice recognition part(12) recognizing the voice by comparing an output signal of a transceiving part(7) of a terminal(100) with previously stored data at a memory(11). |
