基于脉象Hilbert-Huang变换和样本熵的动脉硬化症检测研究  

The Study of the Pulse Signals of Atherosclerosis Based on Hilbert-Huang Transform and Sample Entropy

作　　者：杨成[1] 王学民[1] 孙涛[2] 郁洪强 李想[1] 周鹏[1] 

机构地区：[1]天津大学精密仪器与光电子工程学院,天津300072 [2]第四军医大学生物医学工程学院,西安710032 [3]天津市医疗器械技术审评中心,天津300191

出　　处：《生物医学工程学杂志》2012年第6期1178-1183,共6页Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(51007063)

摘　　要：动脉硬化是一种严重的心血管疾病,对人体危害极大,对其实现早期诊断具有重大意义。本文通过采集健康成年人和动脉硬化患者的脉象,结合Hilbert-Huang变换(HHT)和样本熵分析二者之间的差异。脉象信号经经验模态分解(EMD)分解后,分别计算每个固有模态函数(IMF)分量的样本熵值并对熵值做统计分析,发现动脉硬化患者第一个IMF分量的样本熵值小于正常人,且二者差异有统计学意义。计算HHT边际谱和不同频率段的能量值,发现动脉硬化患者的能量明显向低频移动。计算0～1Hz之间的能量值,发现动脉硬化患者在该频段的能量明显高于正常人,t检验分析结果表明该频段的能量值在二者之间的差异有统计学意义。分析结果表明,通过HHT和样本熵对动脉硬化做出早期诊断是可行的。Atherosclerosis, one of the serious cardiovascular diseases, is very harmful to human bodies. The early di agnosis of arteriosclerosis is of great significance. In this paper, we collected pulse from healthy adults and patients with atherosclerosis. Using HilbertHuang Transform （HHT） and sample entropy, we analyzed the pulse and found the differences between the patients and healthy people. After using the empirical mode decomposition （EMD） to process pulse signals, we calculated sample entropy for each intrinsic mode function （IMF）, and did statistical analy sis of the IMF. The sample entropy of a first IMF from patients with atherosclerosis is less than that from healthy persons, and there was significant differences hetween the healthy and patient groups. In calculating the energy value of different frequencies on the HHT marginal spectrum, we found the energy in patients moved to low frequencies obviously. The energy value of frequency between 01 Hz was significantly higher in patients than in the healthy group. The t test also showed that the values between the two groups had significant differences. The statistics and figures showed that early diagnosis was feasible based on HHT and sample entropy.

关 键 词：脉象 动脉硬化症 Hilbert—Huang变换 样本熵 无创检测 

分 类 号：R318.01[医药卫生—生物医学工程]