Automatic ECG Characteristic Analysis using Wavelet Packet Transform

This paper present the automatic ECG signal analysis by wavelet packet transform (WPT) along with signal squaring, window integration and frame creating to cover each cycle of ECG signal to ease the process to calculate the set on points, and set off points, and time intervals within QRS complexes, T waves and P waves. The results of QRS complex detection from automatic ECG analysis on 15 sets of ECG data from MIT-BIH Arrhythmia with total number of 31,959 beats have shown that selectivity Se = 99.79%, the positive predictivity +P = 99.58% and the detection error DER = 0.632%