公开(公告)号：KR100802984B1

公开(公告)日：2008-02-14

申请号：KR1020060116702

申请日：2006-11-24

Applicant: 연세대학교 산학협력단

Inventor： 강홍구 ,  정기훈 ,  이봉진

IPC: G01D21/00

CPC classification number: G10L15/10 ,  G10L15/02

Abstract: An apparatus and a method for discriminating an unidentified signal using a reference model are provided to analyze characteristic parameters at every time regions and thus to discriminate a sound/music signal in real time. A method of discriminating an unidentified signal using a reference model includes the steps of converting the unidentified sound/music signal into a frequency/time region(S501), extracting characteristic parameters from the unidentified sound/music signal in the frequency/time region(S502), comparing the extracted characteristic parameters with a reference model and discriminating the kind of the unidentified sound/music signal(S503), providing information on the reference model, converting the unidentified sound/music signal into the frequency/time region(S506), and extracting characteristic parameters from the sound/music signal.

Abstract translation: 提供一种用于使用参考模型来识别未识别信号的装置和方法，用于分析每个时间区域的特征参数，从而实时地区分声音/音乐信号。 使用参考模型识别未识别信号的方法包括以下步骤：将未识别的声音/音乐信号转换为频率/时间区域（S501），从频率/时间区域中的未识别的声音/音乐信号中提取特征参数（S502 ），将所提取的特征参数与参考模型进行比较并识别未识别的声音/音乐信号的种类（S503），提供关于参考模型的信息，将未识别的声音/音乐信号转换成频率/时间区域（S506） 并从声音/音乐信号中提取特征参数。