An efficient hilbert envelope and factor analysis based Estimation of R-peaks in ECG signal

Abstract

In the present situation of health informatics, appropriate pre-processing tools are necessary to judge actual condition. In this direction, Electrocardiogram (ECG) is the right tool which shows electrical activity of the heart. This tool gives its output in the form of electrical signal having three different wave components namely; P-wave, QRS wave (complex), and T-wave. In this paper combination of three efficient techniques viz. Digital bandpass filtering (DBPF), Hilbert envelope, and Factor analysis is used for ECG signal analysis. The analysis of ECG signal is done by estimating position of R-peaks in MIT-BIH Arrhythmia (MIT-BIH Arr) database. The proposed technique has proved its utility in cardiology by establishing Sensitivity (Se) of 99.95%, Positive Predictivity (Pp) of 99.97%, Accuracy (Acc) of 99.92% and Signal-to-Noise Ratio (SNR) of 37.71dB in considered 18 datasets of MIT-BIH Arr database. In the common datasets used by different researchers, the proposed methodology secured Se of 99.98%, Pp of 99.98%, Acc of 99.96%, and SNR of 39.78dB. The proposed methodology is also compared with well established research works and showing significant improvements in their parameters viz. Se, Pp and Acc. Different work proposed by various authors on ECG signal analysis have been compared. The proposed technique is definitely important for medical engineering applications in estimating correct health condition (by R-peaks detection).