独立分量分析在室性早搏检测与识别中的应用

Application of Independent Component Analysis in Detection and Identification of Premature Ventricular Contraction

提出用独立分量分析的方法对MIT-BIH心电数据库中的数据进行特征提取,再用改进的BP神经网络进行辨识.实验证明,此方法可以较准确地提取出PVC的特征参数,并能成功地判别正常的心电信号与PVC.

Premature Ventricular Contraction （PVC） is a common arrhythmia phenomenon and has important clinical significance, so it is important to detect and identify the PVC. This paper proposes using Independent Component Analysis to extract features from the data in the MIT-BIH ECG database, then identi- fies PVC with improved BP neural network. Experiment proves that this method can extract feature parameters of PVC more accurately, and succeed in discriminating normal ECG and PVC.