粒子群优化在心电信号稀疏分解中的应用  

Algorithm of particle swarm optimization used in sparse decomposition of electrocardiograms

作　　者：王春光[1] 徐安凤[2] 孙即祥[1] 王红[3] 

机构地区：[1]国防科学技术大学电子科学与工程学院,湖南长沙410073 [2]南阳师范学院计算机系,河南南阳473061 [3]北京跟踪与通信技术研究所,北京100094

出　　处：《计算机工程与设计》2008年第16期4311-4314,共4页Computer Engineering and Design

摘　　要：心电信号的稀疏分解能得到稀疏表示形式,便于心电信号的压缩、波形检测识别等。但稀疏分解计算非常复杂,是一个NP问题。粒子群优化是群体智能优化算法,算法简单易于实现,且搜索效果好,已在很多实际应用中得到验证。把粒子群优化算法用于心电信号稀疏分解的最优匹配原子的搜索,用这种群体智能优化搜索算法代替全局搜索能降低稀疏分解复杂度,同时减少稀疏分解的超完备字典对存储空间的占用。实验结果表明,该方法切实可行。Sparse decomposition of electrocardiograms （ECG） can get sparse representation of ECG, and then next disposal such as compression, wave orientation and identify, can use this sparse representation expediently. But sparse decomposition is very complex （NP problem）. Particle swarm optimization is a kind of optimization algorithm using colony aptitude. Its theory is simple to be realized, and the result of searching is good. These have been proved in some applications. To reduce complexity of sparse decomposition and space of memory, particle swarm optimization is used in the searching the best atom, instead of greedy searching. Examinations have proved this method is feasible.

关 键 词：粒子群优化 稀疏分解 心电信号 字典 匹配追踪 

分 类 号：TN911.72[电子电信—通信与信息系统]