A Novel Deterministic Approach for Automatic Classification of Congestive Heart Failure

T-wave and R-peak irregularities are indicative of electrophysiological abnormalities in the cardiac cycle. These irregularities are key markers in an electrocardiogram (ECG) as they suggest the presence of congestive heart failure (CHF). Specifically, these anomalies reflect altered ventricular repolarization and abnormalities in the depolarization process, which are characteristic of the impaired cardiac function seen in CHF patients. In this work, a deterministic algorithm was developed with the aim of providing a timely and rapid indication of the presence of the pathology, thus allowing early intervention. The algorithm is based on the analysis of features extracted from T-waves and R-peaks in the ECG. The algorithm was tested on two datasets: the Fantasia Database and the BIDMC Congestive Heart Failure Database. After a band-pass filtering (0.5-15 Hz), modified Pan-Tompkins were used to detect R-peaks even in leads where they were negative. Subsequently, T-waves were identified. R-peaks and T-waves amplitudes were used for discriminating healthy and pathological subjects. These features were compared in terms of percentage error (PE), confusion matrix was computed and derivative index, such as True Positive Rate (TPR), Positive Predictive Value (PPV) and Accuracy (ACC), were extracted from it. The T-wave amplitude emerged as the most effective feature achieving a classification accuracy of 100%. The promising results achieved by the developed algorithm indicate a substantial advancement in early detection of the disease during its initial stage. In practical applications, this would allow timely intervention by a physician, thereby enhancing patient care and management.