R-Peak Detection Using Wavelet Scattering Transform for Pre-Term Infant ECG Dataset

Heart-related disease is one of the pre-eminent origins of mortality across the globe. Early and accurate detection of heart-related disease is essential to avoid an uncertain mishap. Peak detection is a pivotal step for the analysis of different spectral signals. The dislocation of peaks disrupts the detection accuracy of a heart-related disease. This study proposes a method for accurate R-peak detection of ECG signal. The acute R-peak detection of ECG signal is achieved using wavelet scattering transform. It eliminates the artifacts associated to signal with the help of inbuilt filters, proves reliable as it identifies the R-peaks at correct location. The wavelet scattering transform generates rational, instructive, and invariable translation representation of signals. The efficiency of the proposed method for R-peak detection is evaluated on pre-term infant ECG dataset available at physionet. The attained average accuracy and specificity for R-peak detection is 99.23% and 99.86% respectively. The proposed method outperforms the state-of-the-art methods namely continuous wavelet transform (CWT) and thresholding technique.