A novel method to model ECG beats using Gaussian functions

ECG signal modeling is an essential prerequisite for detection, classification and compression of ECG signal. In this paper, a novel method is developed to model ECG beats using Gaussian fitting of order eight. First, the baseline is detected from the probability histogram of the ECG signal and each beat is divided into two components according to the one above the baseline and the one below it. Both of them are then modeled separately and incorporated together to construct the entire fit. The MIT-BIH Arrhythmia database has been used for authenticity. The difference between the modeled signal and the original signal is calculated and very low residual error has been found. The RMS error of this method has been determined to be 0.02569, 0.02846, 0.05916, 0.02002 and 0.03169 mV for NSR, APC, PVC, LBBB and RBBB beats respectively.