Fast Algorithm for Real Signal Sparse Decomposition Based on FHT
-
摘要: 提出了一种适合于计算机程序实现的稀疏分解快速算法.该算法利用Gabor原子时频参数结构特征,把信号稀疏分解中计算量极大的内积运算转换成信号的互相关运算,大幅度提高了信号稀疏分解的速度.对于实信号,用快速哈特莱变换(FHT)实现互相关的快速运算.仿真结果证实了该算法的有效性.Abstract: A new signal sparse decomposition approach was proposed for easy software implementation.This algorithm converts very time-consuming inner-products in sparse decomposition into cross-correlations,so that the speed of sparse decomposition is increased greatly.For real signals,cross-correlations are done by fast Hartley transform(FHT).Simulation results verified the effectiveness of the proposed algorithm.
-
Key words:
- sparse decomposition /
- matching pursuit /
- cross-correlation /
- FHT
-
MALLAT S,ZHANG Z.Matching pursuit with time-frequency dictionaries[J].IEEE Trans.on Signal Processing,1993,41(12):3397-3415.[2] 张文耀.基于匹配跟踪的低位率语音编码研究[D].北京:中国科学院研究生院(软件研究所),2002.[3] ARTHUR P L,PHIUPOS C L.Voiced/unvoiced speech discrimination in noise using Gabor atomic decomposition[C]// Prec.of IEEE CASSP.Hong Kong:IEEE Press,2003,I:820-828.[4] DAVIS G,MALLAT S,AVELLANEDA M.Adaptive greedy approximation[J].Coustr.Approx.,1997,13(1):57-98.[5] CANDES E,DONOHO D.New fight frames of curvelets and optimal representations of objects with C2 singularities[R].Standford:Standford University,2002:37.[6] XU Peng,YAO Dezhong.Two dictionaries matching pursuit for sparse decomposition of signals[J].Signal Processing,2006,86:3472-3480.[7] 李恒建,尹忠科,王建英.基于量子遗传算法的图像稀疏分解[J].西南交通大学学报,2007,42(1):19-23.U Hengjian,YIN Zhongke,WANG Jianying.Image sparse decomposition on quantum genetic algorithm[J].Journal of Southwest Jiaotong University,2007,42 (1):19-23.[8] 华泽玺,尹忠科,黄雄华.信号在过完备库上分解中原子形成的快速算法[J].西南交通大学学报,2005,40(3):402-405.HUA Zexi,YIN Zhongke,HUANG Xionghua.Fast atom construction algorithm for signal decomposition in over-complete dictionary[J].Journal of Southwest Jiaotong University,2005,40(3):402-405.[9] Braeewell R N.The fast Hartley transform[J].Proe.of the IEEE,1984,72(8):1010-1018.
点击查看大图
计量
- 文章访问数: 1391
- HTML全文浏览量: 94
- PDF下载量: 444
- 被引次数: 0