Localization of ECG QRS Waves Through Spectral Estimation of Heart Rate
An Electrocardiogram (ECG) is a biomedical signal that contains information about the functioning of the heart which is extracted by means of signal processing to assess cardiac functions.
The main feature of the ECG is the R wave in the QRS complex which defines the basic physiological finding in the heart (i.e.) heart rate. The crucial work in ECG analysis is to detect the R waves which are variant in amplitudes and locations throughout the rhythm strip making it difficult to set an amplitude threshold in priori to detect them. Most of ECG analysis methods rely on the detection of the R waves basically for primary as well as secondary analysis.
In this paper we propose a method to localize the R waves without setting a threshold in priori, nether static nor dynamic. The proposed method firstly achieves an estimate of the heart rate by analyzing the spectrum of the ECG signal in order to determine the set of points that includes the R waves and consequently localize the R waves in the time domain representation of the ECG signal. We estimate the heart rate by detecting the most significant energy frequency component in the spectrum that is in the range 0.5 Hz to 6 Hz and the corresponding time period (i.e. reciprocal of frequency component), then we find the maximum amplitude point in the ECG signal which definitely corresponds to an R wave. Starting from that point we go in both directions till the beginning and the end of the signal at a step size equal to the fundamental period aforementioned (i.e.) estimated heart rate component. As a result, the set that contains expected locations of all R waves is obtained. Finally a symmetrical time domain unity amplitude window is designed which is centered at every expected location found earlier and then the maximum amplitude points’ within the window area are found which are going to represent the exact the R wave locations.
KeywordsECG Detection Heart QRS Rate
- 1.Goutam A, Lee Y-D, Chung W-Y (2008) ECG signal denoising with signal averaging and filtering algorithm. In: Third 2008 international conference on convergence and hybrid information technologyGoogle Scholar
- 2.Shen TW, Tompkins WJ (2005) Biometric statistical study of one-lead ECG feature and body mass index (BMI). In: Proceedings of the 2005 IEEE, engineering in medicine and biology 27th annual conference, Shanghai, China, 1–4 September 2005Google Scholar
- 3.Muthuvel K, Padma Suresh L, Jerry Alexander T (2014) Spectrum approach based classification of ECG signal. In: 2014 international conference on circuit, power and computing technologies [ICCPCT]Google Scholar
- 4.Canan S, Ozbay Y, Karlik B (1998) A method for removing low varying frequency trend from ECG signal. In: 2nd international biometrical engineering daysGoogle Scholar
- 5.MIT-BIH arrhythmia databaseGoogle Scholar