Multimedia search apparatus and method for searching multimedia content using speaker detection by audio data
    63.
    发明授权
    Multimedia search apparatus and method for searching multimedia content using speaker detection by audio data 有权
    多媒体搜索装置及使用音频数据的扬声器检测来搜索多媒体内容的方法

    公开(公告)号:US06317710B1

    公开(公告)日:2001-11-13

    申请号:US09353192

    申请日:1999-07-14

    IPC分类号: G01L1700

    CPC分类号: G10L17/00

    摘要: A multimedia search apparatus and method for searching multimedia content using speaker detection to segment the multimedia content. The multimedia search apparatus receives a search request from a user device. The search request identifies the target speaker for which the search is to be conducted. Based on the search request, the multimedia search apparatus retrieves multimedia content from a multimedia database. The multimedia search apparatus retrieves models, such as Gaussian Mixture Models (GMMs), from a model storage device, corresponding to the target speaker and background data. Based on the retrieved models, the multimedia search device searches the audio data of the multimedia content and segments the audio data. The segments are identified by calculating an average normalized score for a block of frames of the audio data and determining if the average normalized score for the block of frames exceeds one or more predetermined thresholds.

    摘要翻译: 一种多媒体搜索装置和方法,用于使用说话者检测来搜索多媒体内容来分割多媒体内容。 多媒体搜索装置从用户装置接收搜索请求。 搜索请求标识要进行搜索的目标扬声器。 基于搜索请求,多媒体搜索装置从多媒体数据库检索多媒体内容。 多媒体搜索装置从对应于目标说话者和背景数据的模型存储装置中检索诸如高斯混合模型(GMM)的模型。 基于所检索的模型,多媒体搜索装置搜索多媒体内容的音频数据并对音频数据进行分段。 通过计算音频数据的帧块的平均归一化分数并确定帧块的平均归一化分数是否超过一个或多个预定阈值来识别段。