基于稀疏分解的心电数据压缩算法  

The Compression Algorithm for Electrocardiogram Based on Sparse Decomposition

作　　者：王春光[1] 刘金江[2] 孙即祥[1] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073 [2]南阳师范学院计算机系,南阳473061

出　　处：《中国生物医学工程学报》2008年第1期13-17,共5页Chinese Journal of Biomedical Engineering

摘　　要：稀疏分解是近年发展起来的新的信号处理方法,其优势在于分解所用的基(字典)是超完备的,能更真切地反映信号本质,因此能得到信号的稀疏表示,对数据压缩非常有利。利用稀疏分解的这一优势,进行了心电数据压缩的探索研究。通过对MIT-BIH心电数据库中数据的训练学习,构造出的心电数据字典中的原子能反映出心电信号的时频域特点,能用较少的原子重构心电信号。该心电数据压缩算法能够按照实际的要求调整压缩比,且失真较小(压缩比达到20:1时,均方误差只有5.11%)。实验表明,该算法用于心电数据压缩是切实可行的。Sparse decompression is a new theory for signal processing, having the advantage in that the base （dictionary） used in this theory is over-complete, and can reflect the nature of signal. So the sparse decompression of signal can get sparse representation, which is very important in data compression. The algorithm of compression based on sparse decompression is investigated. By training and learning the ECG data in MIT-BIH database, we constructed the over-complete dictionary of ECG.. The atoms in this dictionary accorded with the character of ECG, thus it was possible that a long ECG datum was reconstructed by a few nonzero coefficients and atoms. The proposed compression algorithm could adjust compression ratio according to practical request, and the distortion was low （when the compression ratio was 20： 1, the standard error is 5.11% ）. The experiments have proved the feasibility of the proposed compression algorithm.

关 键 词：稀疏分解 正交匹配追踪 K奇异值分解 心电信号 

分 类 号：TN911.72[电子电信—通信与信息系统]