基于二阶统计量的语音信号时域卷积盲分离算法

张华; 冯大政; 庞继勇

doi:10.3724/SP.J.1146.2009.00574

基于二阶统计量的语音信号时域卷积盲分离算法

doi: 10.3724/SP.J.1146.2009.00574

计量
- 文章访问数: 3917
- HTML全文浏览量: 119
- PDF下载量: 898
- 被引次数: 0
出版历程
- 收稿日期: 2009-04-19
- 修回日期: 2009-11-10
- 刊出日期: 2010-05-19

A Time-Domain Blind Convolutive Separation Algorithm for Speech Signals Based on Second-Order Statistic

摘要

摘要: 该文针对语音信号卷积混迭模型，利用信号间近似不相关和短时平稳等二阶统计特性，根据盲分离固有的排列和尺度不定性，改进最小二乘拟合函数，将关于混迭矩阵的四次函数转化为3组待定参数的二次函数。提出优化该代价函数的非正交联合块对角化算法，分3个子步，每个子步求解一个最小二乘问题，交替估计3组待定参数，逼近代价函数最小点。与ZJBD等同类方法相比，该方法具有计算复杂度低，估计精度高且对初始参数选择不敏感等特点，可直接在时域实现语音卷积盲分离。
- 语音信号处理; 卷积盲分离; 非正交联合块对角化; 二阶统计量
Abstract: For blind convolutive separation of speech signals, a cost function improved from the least squares fitting function is developed based on second-order statistic, which simplifies the quartic-function with respect to the mixture matrix into three quadratic functions. The iternative approach with three sub-steps is proposed to perform the non-unitary joint block-diagonalization. In each sub-step, a closed solution is derived by minimizing the cost function associated with one parameter-group while fixing the others. Furthermore, the feature of low computational complexity is analytically proven. Compared with ZJBD, the simulations illustrate that the proposed algorithm has the merits of robust initialization selection and better estimate accuracy.

HTML全文

参考文献(1)

马晓红, 梁丽丽, 殷福亮. 基于盲源分离理论的麦克风阵列信号有音/无音检测方法[J].电子与信息学报.2007, 29(3):589-592浏览Ma X H, Liang L L, and Yin F L. A voice activity detection method based on blind source separation for microphone array signals[J].Journal of Electronics Information Technology.2007, 29(3):589-592[2]党杰, 林秋华, 殷福亮. 基于盲源分离的多幅顺序图像加密方法[J].电子与信息学报.2007, 29(6):1471-1475浏览Dang J, Lin Q H, and Yin F L. Encryption of multiple sequential images using blind source separation[J].Journal of Electronics Information Technology.2007, 29(6):1471-1475[3]付卫红, 杨小牛, 刘乃安. 基于四阶累积量的稳健的通信信号盲分离算法[J].电子与信息学报.2008, 30(8):1853-1856浏览Fu W H, Yang X N, and Liu N A. Robust algorithm for communication signal blind separation fourth-order- cummulant-based[J].Journal of Electronics Information Technology.2008, 30(8):1853-1856[4]Fadaili E-M, Thirion-Moreau N, and Moreau E. Non-orthogonal joint diagonalization/zeros- diagonalization for source separation based on time-frequency distributions[J].IEEE Transactions on Signal Processing.2007, 55(5):1673-1687[5]徐舜, 刘郁林, 陈绍荣. 一种非平稳卷积混合信号的时域盲分离算法[J].电子与信息学报.2008, 30(3):589-592浏览Xu S, Liu Y L, and Chen S R. A time-domain algorithm for blind source separation of non-stationary convolutive mixtures[J].Journal of Electronics Information Technology.2008, 30(3):589-592[6]Bousbiah-Salah H, Belouchrani A, and Abed-Meraim K. Jacobi-like algorithm for blind signal separation of convolutive mixtures[J].Electronics Letters.2001, 37(16):1049-1050[7]Abed-Meraim K and Beloucharni A. Algorithms for joint block diagonalization[C]. Proc. EUSIPCO04, Vienna, Austria, Sept. 6-10, 2004: 209-212.[8]Ghennioui H, Fadaili E M, and Moreau N T, et al.. A nonunitary joint block diagonalization algorithm for blind separation of convolutive mixtures of sources[J].IEEE Signal Processing Letters.2007, 14(11):860-863[9]易克初, 田斌, 付强. 语音信号处理[M]. 北京: 国防工业出版社, 2001: 146-147.

施引文献

资源附件(0)

访问统计