High Order Statistics For Ventricular Ectopic Beats Detection

Electrocardiogram (ECG) signals processing are vital tools to tracking the cardiac abnormalities, thus it is widely used in assessing and diagnosing the health of the heart. One of the most common cardiac arrhythmias which has been considered by clinicians is ventricular ectopic beats (VEBs), it include premature ventricular contraction (PVC) arrhythmia and others cases. The ECG analysis needs to continuous surveillance of the heartbeat record by cardiologist. To avoid this tedious task for clinicians, several automated diagnostic techniques have been developed in the last years. Most of them depend on the wavelet transform and/or Autoregressive model (AR model) and its various extractive features. In this paper, we are proposing a simple method for heartbeats classification, more precisely, for VEBs arrhythmia detection. It is based on two high order statistical measures (kurtosis, skewness), they are used to discriminate between normal beats (N) and abnormal beats (VEBs), and without the need of mathematical modeling (example: signal transformation such as wavelet, AR model). Obtained results are satisfactory and it proves the potential and efficiency to detect VEB arrhythmia using two high order statistical measures (HOS). Moreover, this method is a simple, fast and easy to implement in several platform for real time practice.