公开(公告)号：US20140207447A1

公开(公告)日：2014-07-24

申请号：US14100884

申请日：2013-12-09

Applicant: Huawei Device Co., Ltd.

Inventor： Hongrui JIANG ,  Xiyong WANG ,  Junbin LIANG ,  Weijun ZHENG ,  Junyang ZHOU

IPC: G10L15/20

CPC classification number: G10L15/20 ,  G10L15/08 ,  G10L25/24 ,  G10L25/84

Abstract: Embodiments of the present invention provide a voice identification method, which includes: obtaining voice data; obtaining a confidence value according to the voice data; obtaining a noise scenario according to the voice data; obtaining a confidence threshold corresponding to the noise scenario; and if the confidence value is greater than or equal to the confidence threshold, processing the voice data. An apparatus is also provided. The method and apparatus that flexibly adjust the confidence threshold according to the noise scenario greatly improve a voice identification rate under a noise environment.

Abstract translation: 本发明的实施例提供一种语音识别方法，包括：获取语音数据; 根据语音数据获取置信度值; 根据语音数据获得噪声场景; 获得对应于噪声场景的置信阈值; 并且如果置信度值大于或等于置信度阈值，则处理语音数据。 还提供了一种装置。 根据噪声场景灵活调整置信阈值的方法和装置大大提高了噪声环境下的语音识别率。

公开(公告)号：US20140207460A1

公开(公告)日：2014-07-24

申请号：US14100822

申请日：2013-12-09

Applicant: Huawei Device Co., Ltd.

Inventor： Hongrui JIANG ,  Xiyong WANG ,  Junbin LIANG ,  Weijun ZHENG ,  Junyang ZHOU

IPC: G10L17/20

CPC classification number: G10L17/20 ,  G10L17/12

Abstract: Embodiments of the present invention provide a voice identification method, including: obtaining voice data; obtaining a first confidence value according to the voice data; obtaining a noise scenario according to the voice data; obtaining a second confidence value corresponding to the noise scenario according to the first confidence value; and if the second confidence value is greater than or equal to a pre-stored confidence threshold, processing the voice data. An apparatus is also provided. The method and apparatus that flexibly adjust the confidence value according to the noise scenario greatly improve a voice identification rate under a noise environment.

Abstract translation: 本发明的实施例提供一种语音识别方法，包括：获取语音数据; 根据语音数据获得第一置信度值; 根据语音数据获得噪声场景; 根据第一置信度获得对应于噪声场景的第二置信度值; 并且如果第二置信度值大于或等于预先存储的置信度阈值，则处理语音数据。 还提供了一种装置。 根据噪声场景灵活调整置信度的方法和装置大大提高了噪声环境下的语音识别率。