ML-DLNN-based Heartbeat Detection System for Diagnosing Cardiac Arrhythmia Disease using ECG Signal

Abstract

A condition when the heart rate is irregular either the beat is too slow or too fast is named Cardiac Arrhythmia (CA). Sudden Cardiac Death (SCD) could be caused by risky arrhythmia disorders. Early detection is required for reducing the fatality rate. In current years, several techniques have been developed for automatic Electrocardiogram (ECG) beat classification. Hence, a novel Memory layer with a Logistic BCM function-based Deep Learning Neural Network (ML-DLNN)-based CA disease detection based on fiducial point detection using ECG signal is proposed. Primarily, by utilizing the First Order Derivation based Kaiser Derived Bessel (FOD-KDB), the noises are removed from the input signal, and the isoelectric line is corrected. Afterward, the waves are detected from the obtained signal using PTA and it is decomposed using Deviation Measure based Ensemble Empirical Mode Decomposition (DM-EEMD). The fiducial point is accurately detected from the decomposed signal using CPA. From the fiducial point, the relationship betwixt each interval is computed and the features are extracted. Next, by utilizing Frechet Distribution-based Lemurs Optimization Algorithm (FD-LOA), the important features are selected. For classifying the conditions, both of them are given into the ML-DLNN. The proposed mechanism’s experimental outcomes are contrasted with the conventional approaches, which exhibit the proposed one’s higher detection accuracy