Deep Learning Approach for Coronary Artery Disease Detection and Severity Analysis Using ECG Signals

Globally, cardiovascular disease is one of the main causes of death. It is caused by the narrowing or obstruction of the coronary arteries due to the buildup of fatty substances. Electrocardiogram(ECG) signals can immediately assess a patient's heart health. But ECG perturbations may be misinterpreted during the manual interpretation or by using conventional techniques in ECG equipment. Implementation of Deep Learning techniques for identification of Coronary Artery Disease(CAD) using ECGs has proven more accuracy and are less time consuming. Deep learning algorithms can learn complicated patterns and temporal correlations straight from sequential data, making them ideal for ECG waveform analysis. This paper proposes a hybrid deep learning model for CAD detection using ECG signals. The hybrid deep learning presented in the paper incorporates advanced learning techniques such as Bidirectional Long short-term memory (BI-LSTM) and Neural Basis expansion analysis for interpretable time series (N-BEATS). The proposed model has the capacity to serve as a valuable diagnostic tool for categorising typical ECG patterns to diagnose Coronary disease and identify of severity level of disease.