计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (36): 34-36.DOI: 10.3778/j.issn.1002-8331.2008.36.009

• 博士论坛 • 上一篇    下一篇

似然比测试语音激活检测算法研究

雷建军1,王 坚2,刘 刚3,郭 军3   

  1. 1.天津大学 电子信息工程学院,天津 300072
    2.中央财经大学 信息学院,北京 100081
    3.北京邮电大学 信息工程学院,北京 100876
  • 收稿日期:2008-09-18 修回日期:2008-10-31 出版日期:2008-12-21 发布日期:2008-12-21
  • 通讯作者: 雷建军

Research on voice activity detection algorithm based on Likelihood Ratio Test

LEI Jian-jun1,WANG Jian2,LIU Gang3,GUO Jun3   

  1. 1.School of Electronic Information Engineering,Tianjin University,Tianjin 300072,China
    2.School of Information,Central University of Finance and Economics,Beijing 100081,China
    3.School of Information Engineering,Beijing University of Posts and Telecommunications,Beijing 100876,China
  • Received:2008-09-18 Revised:2008-10-31 Online:2008-12-21 Published:2008-12-21
  • Contact: LEI Jian-jun

摘要: 基于似然比测试的语音激活检测算法基于假设检验原理,引入了对噪声的降噪处理,具有良好的噪声鲁棒性,算法高效且易于实现。各种改进算法针对语音模型、判决规则等核心问题进行了系统研究,检测性能得到了进一步提高。对此类算法进行了系统的论证,结合实验,分析比较了它们的性能差异和各自的优缺点,并指出了进一步的研究方向。

关键词: 语音激活检测, 似然比测试, 语音模型, 判决规则

Abstract: Voice activity detection algorithm based on Likelihood Ratio Test(LRT) using the principle of hypothesis testing,introducing the reduction of noise,has good noise robustness,and it is efficient and easy to achieve.Recent researches on speech model and decision rule improve the performance of voice activity detection.In this paper,the basic and improved likelihood ratio test algorithms are demonstrated.Then,their respective advantages and disadvantages of the approaches are outlined.Finally,the challenges that are still worth to further research are also pointed out.

Key words: voice activity detection, Likelihood Ratio Test(LRT), speech model, decision rule