基于统计模型的麦克风阵列语音增强算法  

MICROPHONE ARRAY SPEECH ENHANCEMENT ALGORITHM BASED ON STATISTICAL MODELS

作　　者：涂井先 冀占江 覃桂茳 蒲保兴 Tu Jingxian;Ji Zhanjiang;Qin Guijiang;Pu Baoxing(School of Data Science&Software Engineering,Wuzhou University,Wuzhou 543002,Guangxi,China)

机构地区：[1]梧州学院大数据与软件工程学院,广西梧州543002

出　　处：《计算机应用与软件》2024年第11期335-340,共6页Computer Applications and Software

基　　金：广西自然科学基金项目(2018JJB170034,2020JJA110021,2020GXNSFAA238013);梧州学院校级科研项目(2017B004)。

摘　　要：提出一种基于统计模型的麦克风阵列语音增强算法。为了估计第一通道语音信号谱幅度平方,假设第一通道语音信号谱的实部和虚部相互独立,并服从方差相同的高斯分布。先利用贝叶斯公式计算第一通道语音信号谱幅度平方的后验概率,再利用数学期望的计算公式得到第一通道语音信号谱幅度平方的最小均方误差估计,最后用第一通道观测语音信号的谱角度来估计第一通道语音信号的谱角度。实验结果表明,所提出的算法优于三种传统的语音增强算法。This paper presents a microphone array speech enhancement algorithm based on statistical models.In order to estimate the magnitude-squared spectrum(MSS)of the first channel speech signal,it was assumed that the real part and imaginary part of the first channel speech signal spectrum were independent of each other and obeyed the Gaussian distribution with the same variance.The posterior probability of the MSS of the first channel speech signal was calculated by using the Bayesian formula,and the minimum mean square error(MMSE)estimation of the MSS of the first channel speech signal was obtained by using the mathematical expectation formula,and finally the spectrum angle of the first channel speech signal was estimated by using the spectrum angle of the first channel observed signal.The experimental results show that the proposed algorithm is superior to the three traditional speech enhancement algorithms.

关 键 词：语音增强 最小均方误差估计 去噪性能 

分 类 号：TP302.7[自动化与计算机技术—计算机系统结构]