计算机技术与发展2016，Vol.26Issue(3)：71-74,4.DOI:10.3969/j.issn.1673-629X.2016.03.017
低信噪比环境下改进的语音端点检测算法
Improved Speech Endpoint Detection under Low SNR Environment
摘要
Abstract
Endpoint detection has a very important role in speech recognition,its accuracy will directly affect the accuracy of speech recog-nition system. In order to improve the accuracy of speech endpoint detection under low SNR environment,an endpoint detection algorithm based on spectral subtraction of multitaper spectrum estimation and spectral entropy is proposed. Firstly,it uses improved spectral subtrac-tion of multitaper spectrum estimation to denoise speech signal in order to improve signal to noise ratio,and then it utilizes energy-entro-py-ratio algorithm to make endpoint detection for speech signal denoised. Simulation experiment results show that compared with com-mon endpoint detection algorithm,this algorithm has good endpoint detection accuracy and certain robustness in low SNR environment. It proves the effectiveness of the proposed algorithm.关键词
多窗谱估计/改进谱减法/谱熵/语音增强/端点检测Key words
multitaper spectrum estimation/improved spectral subtraction/spectral entropy/speech enhancement/endpoint detection分类
信息技术与安全科学引用本文复制引用
董胡..低信噪比环境下改进的语音端点检测算法[J].计算机技术与发展,2016,26(3):71-74,4.基金项目
湖南省科技计划项目(2012FJ3010) （2012FJ3010）
湖南省教育科研项目(12C0952) （12C0952）
长沙师范学院科研基金项目(XXZD201218) （XXZD201218）
长沙师范学院大学生研究性学习和创新性实验计划项目(DXYC201510) （DXYC201510）
