An Efficient Electrocardiogram R-peak Detection Exploiting Ensemble Empirical Mode Decomposition and Hilbert Transform

Abstract

The electrocardiogram (ECG) waveforms, that are P-, Q-, R-, S-, and T-waves expressing the heart activities, have been widely employed for the detection of heart disasters using the distance between adjacent peaks. Among them, R-peak plays the most important role in diagnosing heart diseases. In this work, we propose an efficient R-peak detection solution that utilizes Butterworth bypass filter, Ensemble Empirical Mode Decomposition (EEMD), and Hilbert Transform (HT) for ECG signals. In our approach, EEMD is employed to extract QRS complexes in ECG signals while Hilbert transform is then applied for obtaining the envelope for the R-peak detection of the ECG signal. Firstly, the baseline wander, artifacts, and noises of raw ECG signals will be removed by using a Butterworth filter. The filtered signal is decomposed into a set of Intrinsic Mode Functions (IMFs), monocomponent signals, by implementing the Ensemble EMD method and the first three IMFs that carry sufficient R-peak information are then combined. After that, the first derivative of the combined signal is calculated to figure out the minima or maxima points and subsequently, the differentiated signal will be transformed to determine the envelope by using HT. Finally, based on that, the maximal positions which describe the R-peak positions are marked. Numerical experiments have been done on a popular public database, MIT-BIH Arrhythmia Database, for verifying the performance of our proposed solution in comparison with conventional algorithms. The obtained results prove that our developed approach outperforms the comparative conventional ones. It achieves the average sensitivity and specificity of 98.74% and 98.71% respectively with a detection error rate of 0.028%.