一种基于小波和时频分解的端点检测方法  

A Method of Endpoint Detection Based on Wavelet and Time-frequency Decomposition

作　　者：董胡[1] 钱盛友[1] 

机构地区：[1]湖南师范大学物理与信息科学学院,长沙410081

出　　处：《微计算机信息》2007年第30期228-229,264,共3页Control & Automation

基　　金：湖南省自然科学基金(02JJY2059)

摘　　要：提出了一种新的基于小波和时频分解的语音端点检测方法。首先通过小波分解对含噪信号进行增强,然后采用Matching pursuits算法对去噪信号进行时频分解,使得信号在时频平面上具有较明显的魏格纳能量分布,最后利用该特点设定合适的门限来进行语音端点检测。实验结果表明,该方法对低信噪比的语音端点检测仍有效。A new method of speech endpoint detection based on wavelet and decomposition in time-frequency domain is proposed. Signal containing noise is strengthened through wavelet decomposition, then the reconstructed signal is decomposed by Matching pursuits algorithm in time-frequency domain to obtain apparent Wigner energy distribution. Finally, a suitable threshold is selected to detect the speech endpoint according to these properties. Experimental results demonstrate this approach is effective for low SNR speech endpoint detection.

关 键 词：小波 语音增强 时频分解 端点检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]