음성 인식 시스템 및 방법
    11.
    发明公开
    음성 인식 시스템 및 방법 无效
    用于识别语音的系统和方法

    公开(公告)号:KR1020090035222A

    公开(公告)日:2009-04-09

    申请号:KR1020070100368

    申请日:2007-10-05

    Abstract: A voice recognition system and a method thereof for restoring the voice of the high quality are provided to maintain the performance of voice recognition by a voice recognition server by transmitting voice property parameter and the voice encoding parameter encoding the voice property parameter into the natural disposition. A voice property extracting part(110) outputs voice property parameter by extracting voice property parameter from voice signal inputted from a terminal(100). A voice signal part(130) outputs the terminal is encoded and the voice encoding parameter for the voice restore is outputted. A packet generating part(150) packetizes a packet including voice property parameter and voice encoding parameter. A packet transmission part inputs the packet and transmits the packet to a voice recognition server through the communications network.

    Abstract translation: 提供了一种用于恢复高质量语音的语音识别系统及其方法,通过将语音属性参数和语音编码参数传送到自然配置中来保持语音识别服务器的语音识别性能。 语音特征提取部分(110)通过从从终端(100)输入的语音信号中提取语音属性参数来输出语音属性参数。 语音信号部分(130)输出端被编码,并且语音恢复的语音编码参数被输出。 分组产生部分(150)对包括语音属性参数和语音编码参数的分组进行打包。 分组传输部分输入分组,并通过通信网络将分组发送到语音识别服务器。

Patent Agency Ranking