基于小波变换和K-means聚类算法的心电信号特征提取  

Feature Extraction of Electrocardiogram Signal Based on Wavelet Transform and K-means Clustering Algorithm

作　　者：王瑞荣[1] 余小庆[1] 朱广明[1] 王敏[2] Wang Ruirong Yu Xiaoqing Zhu Guangming Wang Min(Coilege of Life Information Science ＆ In- strument Engineering, Hangzhou Dianzi University, Hangzhou-Zhejiang 310018, China)

机构地区：[1]杭州电子科技大学生命信息与仪器工程学院,浙江杭州310018 [2]杭州红十字会医院,浙江杭州310003

出　　处：《航天医学与医学工程》2016年第5期368-371,共4页Space Medicine & Medical Engineering

基　　金：国家自然科学基金(61374005)

摘　　要：目的研究一种基于小波变换和K-means聚类算法的心电信号特征提取方法,根据特征点信息判断心电是否正常。方法利用小波变换和形态学滤波方法去除工频干扰、肌电干扰和基线漂移等主要的噪声之后,利用K-Means聚类算法提取出心电信号的QRS波群,P波和T波这3个主要的特征点,实现心电智能诊断。结果实验数据取自MIT-BIH数据库,多次实验结果显示QRS波群的阳性检测度(P+)达到99.68%和灵敏度(Se)达到99.21%,P波和T波的检测准确度分别达91.43%和97.01%。结论相对于其它方法,本文心电特征提取方法准确度较高,具有一定参考价值;在移动医疗和临床医疗方面具有一定实用价值。Objective To study a method to extract the features of ECG signal based on wavelet transform and K-means clustering algorithm so as to provide useful information for detection of cardiac disease or abnormality. Methods After removing the power frequency interference,electromyography（ EMG） interference and baseline drift with wavelet transform and morphological filtering methods,the K-Means clustering algorithm was adopt to extract the QRS complex,P wave and T wave,and the characteristics of the ECG signal,thus the automatic delineation of ECG was realized. Results The experimental data was taken from the MIT-BIH database. The results showed that the detection rate of the positive（ P＋） was 99. 68% and sensitivity（ Se） of the QRS complex was 99. 21%. While the extraction precision of P and T waves were 91. 43% and 97. 01% respectively. Conclusion The method in this paper has higher accuracy as compared with some other methods in ECG feature extraction,and it can provide reference for other researchers,and is practical in mobile medical and clinical.

关 键 词：心电信号 小波变换 K-MEANS QRS波群 P波 T波 

分 类 号：R540.41[医药卫生—心血管疾病]