| 摘要 |
A voice recognition device and a method thereof are provided to accurately recognize an inputted user voice by effectively selecting the vocabulary set for searching. A first decoding unit(100) produces a unidentified pronunciation sequence by decoding the voice inputted by the user. A transceiver(110) transmits the produced pronunciation sequence to a server(20). The server extracts a filtered vocabulary set from a bulk vocabulary set. A second decoding unit(120) selects a candidate vocabulary group which is most similar to the unidentified voice. |
