ECG Arrhythmia Classification Using Recurrence Plot and ResNet-18

Cardiovascular diseases are the leading cause of death worldwide, claiming approximately
 17.9 million lives each year. In this study, a novel CAD system to detect and classify electrocardiogram (ECG) signals is presented. Designed employs recurrence plot (RP) approach that transforms ECG signal into 2D representative colour image, finally performing their classifications via employment Deep Learning architecture (ResNet-18). Novel includes two steps, where first step preprocessing one, which performs segmentation data two-second intervals, forming images RP approach; following, in second step, classified by ResNet- 18 network. The proposed method evaluated on MIT-BIH arrhythmia database 5 principal types arrhythmias have medical relevance should be classified. can before-mentioned quantity according AAMI Standard appears demonstrate good performance terms criteria: overall accuracy 97.62%, precision 95.42%, recall F1-Score 95.06%, AUC 95.7% competitive with better state-of-the-art systems. Additionally. demonstrated ability mitigating problem imbalanced samples.