Detection of PVC by using a wavelet-based statistical ECG monitoring procedure

Automatic detection of premature ventricular contractions (PVCs) is essential to timely diagnosis of dangerous heart conditions. However, accurate detection of PVCs is challenging because of multiform PVCs. In this paper, an electrocardiographic (ECG) monitoring procedure based on wavelet-based statistical process control is proposed for diagnosing PVC beats. After ECG signals are decomposed and denoised via discrete wavelet transforms, significant wavelet coefficients are extracted through a sparse discriminant analysis for constructing a monitoring statistics based on Hotelling's T2 statistics. The proposed monitoring method alarms when the monitoring statistics exceeds the predetermined upper control limit. We demonstrated in this study the effectiveness of the proposed procedure by using the MIT-BIH arrhythmia database: the accuracy, sensitivity, specificity, and positive predictivity were obtained as 0.979, 0.872, 0.988, and 0.846, respectively.