基于ARMA模型的ECG分类和压缩

Cardiac arrhythmias classification and compression based on ARMA model

心电信号(ECG)对医生诊断心脏疾病极为重要。现存许多ECG分类技术存在实现困难、处理时间长和只能对2～3类ECG进行分类的不足。本文介绍了计算简单的ARMA模型的ECG分类法,利用ARMA模型系数作为特征对ECG信号进行分类和压缩。在对信号特征分类时,采用了非线性二次判别函数的形式。利用文中方法对MIT-BIH标准数据库中NSR、APC、PVC、SVT、VT和VF各200个样本信号进行测试,获得了94.28%～99.28%的分类精度。

Electrocardiogram (ECG) signal is important for physician to diagnose cardiac diseases. Various existing techniques on ECG classification have been reported. Generally, these techniques classify two or three arrhythmias only or have significantly large processing times. A simpler autoregressive-moving average (ARMA) based technique is proposed to classify ECG for diagnosis in this paper. The ARMA coefficients have been used to classify and compress ECG directly. ARMA technique has been used for classification into arrhythmias such as normal sinus rhythm (NSR), premature ventricular contraction (PVC), atrial premature contraction (APC), ventricular tachycardia (VT), ventricular fibrillation (VF), and superventricular tachycardia(SVT). Two hundred data samples from each of them have been utilized to classify and test. The accuracy of detecting these arrhythmias proposed are 93.5% to 97.86% using the quadratic discrimination function stage-by-stage.\;