基于近似KLT域的语音信号压缩感知

郭海燕; 杨震

doi:10.3724/SP.J.1146.2008.01704

基于近似KLT域的语音信号压缩感知

doi: 10.3724/SP.J.1146.2008.01704

郭海燕,
杨震

基金项目:

国家863计划重点项目(2006AA010102)国家自然科学基金(60971129)和江苏省普通高校研究生科研创新计划项目(CX0913-148Z)资助课题

计量
- 文章访问数: 3470
- HTML全文浏览量: 82
- PDF下载量: 1054
- 被引次数: 0
出版历程
- 收稿日期: 2008-12-15
- 修回日期: 2009-05-18
- 刊出日期: 2009-12-19

Compressed Speech Signal Sensing Based on Approximate KLT

摘要

摘要: 压缩感知是近年来兴起的研究热点，该文基于语音信号在KLT域的稀疏特性，提出了基于模板匹配的近似KLT，并在基于模板匹配近似KLT域上研究了语音信号的压缩感知性能。首先验证语音信号在基于模板匹配近似KLT域上的稀疏性，然后由语音信号与观测矩阵构造相应的观测，采取固定分配每帧观测个数和按帧能量自适应分配每帧观测个数两种方案，再以观测为已知条件利用L1优化算法重构语音信号在基于模板匹配近似KLT域的稀疏系数向量，进而重构原始语音信号。实验表明，语音信号在基于模板匹配的近似KLT域的压缩感知性能较好。
- 语音合成; 压缩感知; 稀疏性; L1优化; Karhunen-Loeve变换(KLT)
Abstract: Compressed Sensing is a research focus rising in recent years. On the basis of the signals sparse representation in the KLT domain, this paper proposes an approximate KLT method using template matching and studies on the corresponding compressed speech signal sensing. First, it verifies the sparsity of speech signal in the approximate KLT domain. Second, by speech signal and a measurement matrix, it arranges measurements of fixed or adaptive length according to frame energy. Third, according to the measurements, it finds the speech signals sparsest coefficient vector through L1 optimization algorithm to recover the speech signal. Simulation results demonstrate that compressed speech signal sensing in the approximate KLT using template matching has good performance.

HTML全文

参考文献(1)

Donoho D. Compressed sensing[J].IEEE Transactions onInformation Theory.2006, 52(4):1289-1306[2]Tsaig Y and Donoho D. Extensions of compressed sensing[J].Signal Processing.2006, 86(3):533-548[3]Cands E, Romberg J, and Tao T. Robust uncertaintyprinciples: Exact signal reconstruction from highlyincomplete frequency information[J].IEEE Transactions onInformation Theory.2006, 52(2):489-509[4]Romberg J. Imaging via compressive sampling. IEEE SignalProcessing Magazine, 2008, 25(2): 14-20.[5]Duarte M, Davenport M, Takhar D, Laska J, Sun T, Kelly K,and Baraniuk R. Single-pixel imaging via compressivesampling[J].IEEE Signal Processing Magazine.2008, 25(2):83-91[6]Gemmeke J F and Cranen B. Using sparse representations formissing data imputation in noise robust speech recognition.European Signal Processing Conf. (EUSIPCO), Lausanne,Switzerland, August 2008: 987-991.[7]Baraniuk R G. Compressive sensing. IEEE Signal ProcessingMagazine, 2007, 24(4): 118-121.[8]Baron D, Wakin M, Duarte M, Sarvotham S, and Baraniuk R.Distributed compressed sensing. Technical Report ECE-0612,Electrical and Computer Engineering Department, RiceUniversity, December 2006.[9]Chen S S, Donoho D L, and Saunders M A. Atomicdecomposition by basis pursuit[J].SIAM Review.2001, 43(1):129-159[10]A帕普里斯, SU佩莱著, 保铮, 冯大政, 水鹏朗译. 概率、随机变量与随机过程. 第4 版, 西安: 西安交通大学出版社,2004: 209-213.[11]Coifman R R and Donoho D L. Translation-InvariantDe-noising. New York: Springer-Verlag, 1995: 125-150.

施引文献

资源附件(0)

访问统计