R-peak detection using a complex fractional wavelet

Abstract

In a recent paper we have introduced and used a real fractional wavelet for the detection of the QRS complex in ECG signal. This fractional wavelet is the second derivative of the Cole-Cole distribution which is widely used in the modelling of dielectrics by a fractional power pole. This work presents a QRS detection scheme using a complex fractional wavelet function. The absolute value of this wavelet is the Cole-Cole distribution function. The method consists of using the real and the imaginary parts of the CWT along with the multiscale product to distinguish the R waves from the high P and T waves, the noise and the artefacts. The study of the proposed algorithm and the analysis of the results have been done using the first four-minutes of the 24 records of the MIT-BIH arrhythmia database. The R wave detection results have been compared to the annotation file accompanying each ECG signal. We have found that the proposed detection algorithm gives high accuracy.