摘要
基于Frank正交心电导联早期心肌梗死的特征提取和分类,提出利用多通道回归(multivariate autoregressive,MAR)模型对心电信号(electrocardiogram,ECG)进行建模,以MAR系数为心电特征,对PTB诊断数据库中的正常状态病人、早期心肌梗死和急性期心肌梗死进行分类测试.结果表明,利用该方法从Frank心电导联中提取特征对早期心肌梗死和急性期心肌梗死进行分类诊断是可行的,分类精度能获得有效提高.
Most of existing myocardial infarction （MI） techniques focus on the detection of acute myocardial infarction using standard electrocardiogram （ECG） leads. The study of myocardial infarction in early stage （MIES） feature extraction was performed using Frank orthogonal leads. Multivariate autoregressive （MAR） model coefficients were used as ECG features for the classification. The data in the analysis including health control （ HC）, MIES and AMI were collected from FFB diagnostic ECG database. The experimental results show that it is feasible to separate MIES from AMI using Frank orthogonal leads. The accuracy pared with that of standard ECG leads using the same feature of detecting MIES and AMT increased effectively com- dimension and threshold value of zero.
出处
《深圳大学学报（理工版）》
EI
CAS
北大核心
2009年第2期137-142,共6页
Journal of Shenzhen University(Science and Engineering)
基金
浙江省自然科学基金资助项目(Y104284)
关键词
心肌梗死
心电图
Frank导联
特征提取
myocardial infarction
electrocardiogram
Frank leads
feature extraction
