针对语音信号的弱稀疏性，提出一种新的基于混合矩阵估计的欠定语音盲分离方法。
This paper proposes a new method based on mixing matrix estimation for underdetermined blind speech separation, aiming at speech signals under weak sparseness.
该方法利用Curvelet多尺度几何分析后信号的稀疏性特点，采用了C - means聚类方法寻求混合矩阵估计值，把该估计值作为算法初始值。
According to signals sparsity by Curvelet transform, the mixed matrix can be estimated with C-means cluster analysis, and the estimated value is looked as initial value of BSS algorithm.
利用稀疏分量的直线聚类性，提出了欠定盲源分离中估计混合矩阵的一种方法。
A method of the mixing matrix estimation in underdetermined source separation is proposed, which is based on the linear clustering of sparse component.
实验分为两个过程：（1）估计混合矩阵；
The experiment has two steps:(1)estimating the mixing matrix;
该方法采用固定点ica算法来估计多径信道的混合矩阵，从而提取信道的延迟信息。
The mixture matrix of multi-path channel can be estimated using a fast fixed-point algorithm, and then the delay information of channel can be obtained.
对四阶累量混合波达方向矩阵进行特征分解，可实现有色高斯噪声背景中空域信号二维空间谱估计。
We can estimate two dimensional spatial spectra of sources in colour Guassian noises by eigen decomposing the matrix.
这些时刻的观测信号矢量就是对混合矩阵中与该源信号对应的列矢量的估计，利用这一性质可以估计出混合矩阵。
The vectors of observation at these instants, are the estimate of the corresponding columns at the mixing matrix.
在估计出混合矩阵的基础上，利用最短路径法分离出源信号。
Then, the source signals can be recovered by the shortest path method.
考虑含有两个方差分量矩阵的多元混合模型，将一元混合模型下的谱分解估计推广到多元模型下。
Spectral decomposition estimators of variance component matrix in mixed linear model are generalized to multivariate mixed linear model.
考虑含有两个方差分量矩阵的多元混合模型，将一元混合模型下的谱分解估计推广到多元模型下。
Spectral decomposition estimators of variance component matrix in mixed linear model are generalized to multivariate mixed linear model.
应用推荐