Efficient R peak Detection Algorithm from ECG using Combination Stationary Wavelet Transform and Hilbert Transform

Authors

  • Manoj Kumar Ojha, Dr.S.Wadhwani, Dr. A.K.Wadhwani, Dr. AnupamShukla

Abstract

Detection of early and accurate R peaks is important to analyze electrocardiogram (ECG) signals in many critical situations. Noises and the interferences affect the quality of ECG signals and hence performance or end result. This poses great challenges in the analysis. Existing method of R peak detection and analysis are not adequate. So, it is obligatory to design an advanced technique that preprocesses and detects R peaks effectively. This novel algorithm increases efficiency and add to quality of detection. The novelty of this paper is that an algorithm is that the proposed algorithm accurately detects peaks, missed and fast R peaks. The algorithm is a combination of Stationary Wavelet Transform (SWT) and Hilbert Transform (HT). HT has the ability to separate the positive peaks. It envelops the peak signal and helps in finding the R peaks in the high probability locations. On the other hand, SWT has the ability to divide the ECG components into its various sub bands that are located in specific time-frequency; it can eliminate the false R peaks and locate the true R peaks. To analyze and identify thresholds for R peaks in ECG signal is used as a reading template. By changing the heart rate and quality of the signal, thresholds are reset and modified every three seconds. The algorithm is tested and verified according to Positive Predictivity Value (PPV) and Sensitivity Value (SEV) in PhysoNet data sets. The R peaks yield of 99.93% SEV and 99.89% PPV were obtained. Identification of false R peaks are also largely suppressed and the real R peaks are identified with high accuracy thus indicating algorithm’s excellent performance and accuracy.

Keywords: ECG Signal, Signal Processing, Hilbert Transform, and Stationary Wavelet Transform. 

Published

2020-12-31

Issue

Section

Articles