数据采集与处理Issue(2)：359-364,6.DOI:10.16337/j.1004-9037.2015.02.014
基于隐马尔可夫模型的非监督噪声功率谱估计
Unsupervised Noise Power Estimation Using Hidden Markov Model
摘要
Abstract
Noise estimation is a fundamental part of speech enhancement.Most traditional methods are heuristic which can not enable the optimal estimation.An unsupervised noise power estimation is presen-ted based on maximum likelihood.A log-power statistical model is constructed using hidden Markov model (HMM)in each subband.This model comprises speech and nonspeech Gauss components,and the mean value of nonspeech Gauss component is the estimation of noise power.Moreover,speech may be long-term absent,some constraints are introduced to this model for stability.The experiments vali-date that the proposed method can obtain the maximum likelihood noise estimation and outperforms con-ventional heuristic methods.关键词
语音增强/噪声功率谱估计/隐马尔可夫模型/极大似然准则/模型约束Key words
speech enhancement/noise power estimation/hidden Markov model/maximum likelihood criterion/model constraints分类
信息技术与安全科学引用本文复制引用
许春冬,战鸽,应冬文,李军锋,颜永红..基于隐马尔可夫模型的非监督噪声功率谱估计[J].数据采集与处理,2015,(2):359-364,6.基金项目
国家重点基础研究发展计划(“九七三”计划)(2013CB329302)资助项目 （“九七三”计划）
国家自然科学基金(61271426,10925419,90920302,61072124,11074275,11161140319)资助项目 （61271426,10925419,90920302,61072124,11074275,11161140319）
中国科学院战略性先导科技专项(XDA06030100,XDA06030500)资助项目 （XDA06030100,XDA06030500）
中国科学院重点部署(KGZD-EW-103-2)资助项目 （KGZD-EW-103-2）
江西理工大学科研基金(NSFJ2015-G21)资助项目。 （NSFJ2015-G21）
