The Coronary Atherosclerosis Detection: A Diagnostic Approach of Digital Electrocardiography Image Interpretation

Coronary Atherosclerosis is recognized as the predominant cause of Chronic Heart Failure. The standard diagnostic modality for Coronary Atherosclerosis involves echocardiography (ECO); however, the accessibility of such diagnostic equipment remains constrained within several healthcare facilities. Consequently, there is a demand for a more straightforward diagnostic approach, specifically through the utilization of Electrocardiography (ECG). This investigation endeavors to introduce an automated methodology for the detection of Coronary Atherosclerosis, enhancing the precision of ECG interpretation via digital imaging predicated on cardiac signal detection graphs obtained through this modality. The algorithm employed to delineate the ECG signal dynamics for the diagnostic identification of Coronary Atherosclerosis is the Continuous Wavelet Transform (CWT). The application of CWT follows the execution of multiple pre-processing phases on the ECG signal graph, including image cropping to isolate the graph within the ECG and segmentation to eliminate background interference, thereby concentrating on the graph for subsequent analysis. This preprocessing facilitates the interpretative process through the CWT algorithm. The analysis proceeds with a classification phase employing a Convolutional Neural Network (CNN), which differentiates between normal patient conditions and those presenting with Coronary Atherosclerosis and normal ECG. The experimental outcomes derived from the application of the proposed algorithm demonstrate a significant enhancement in the accuracy of ECG interpretation, achieving a diagnostic accuracy rate of 96% for the detection of Coronary Atherosclerosis.