基于循环相关的扩展空间解相关算法及其模式检测中的应用  

Extended Spatial Decorrelation Algorithm Based-on Cycle Correlation And Its Application in Optical Imaging of Functional Brain

作　　者：黄晓斌[1] 万建伟[1] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073

出　　处：《信号处理》2006年第2期233-237,共5页Journal of Signal Processing

摘　　要：针对空域扩展空间解相关(spatial extended spatial decorrelation;SESD)算法对参数选择鲁棒性差的缺点,使用一维循环相关操作代替原二维平移相关操作,提出了一种基于循环相关的扩展空间解相关(extended spatial decorrelation based- on cycle correlation;ESDBCC)算法。在无噪和有噪环境下,研究了SESD算法与ESDBCC算法对参数选择的鲁棒性,结果表明ESDBCC算法有以下两个显著优点:1)与二维平移相关操作相比,一维循环相关操作更适合计算机计算,从而降低了计算量;2)在有噪和无噪环境下,ESDBCC算法对参数选择的鲁棒性都要明显好于SESD算法。To the shortage of poor robustness for the parameter selection in the spatial extended spatial decorrelation（SESD） algorithm, substitute the old two-dimension correlation manipulation for the new one-dimension cycle correlation manipulation, and present a new ESD algorithm--the extended spatial decorrelation algorithm based-on cycle correlation（ ESDBCC ） algorithm. In the noiseless and noise circumstances, study the robustness for the parameter selection of SESD and ESDBCC algorithms, the results show that our ESDBCC algorithm has the following two remarkable merits： 1 ） one-dimension cycle correlation manipulation is more suitable to the computer operation, so the computational burden is relieved; 2） robustness for the parameter selection of our ESDBCC algorithm is obviously better than the SESD algorithm.

关 键 词：扩展空间解相关 循环相关 光学成像 

分 类 号：TN914[电子电信—通信与信息系统] TP273[电子电信—信息与通信工程]