R peak detection in electrocardiogram signal based on a combination between empirical mode decomposition and Hilbert transform

Abstract

In this paper, we present a combination between Empirical Mode Decomposition (EMD) approach and Hilbert transform approach for the purpose of R peak detection in Electrocardiogram (ECG) signal. This algorithm uses the EMD to find the signal which highlights the region of the QRS complex in ECG signal by combining the first three IMF that contain sufficient information about the region of the QRS complex then the envelope obtained from Hilbert transform to detect the R-peaks. The proposed method requires the following stages: eliminate the baseline wander from the original ECG signal, decompose the resulting filtered ECG signal into a collection of AM-FM components called Intrinsic Mode Functions (IMF) which are obtained by using Empirical Mode Decomposition approach, sum the first three Intrinsic Functions Mode (IMFs) which contain enough information about the QRS complex, calculate the first derivative of the sum signal to get the points of minima or maxima, The differentiated signal is then transformed using Hilbert transform and then we determine the Hilbert envelope, and finally, find the positions of the maximum which represent the positions of the R peaks. The proposed algorithm is evaluated by using the ECG MIT-BIH database and is compared to another technique, proposed by Taouili. The performance of the algorithm is confirmed by a sensitivity of 94.71 %, compared to Se=91.17 given by Taouli's method.