Recognition of cardiac abnormalities from synchronized ECG and PCG signals

Phonocardiogram signals (PCG) and electrocardiogram signals (PCG) have been used separately for decades to diagnose heart abnormalities. Combining these two synchronous signals is expected to enhance the diagnosis for better medical management of patients. This paper's objective is to highlight the performance comparison between the diagnosis of heart abnormalities based only on PCG recordings and that based on synchronized PCG and ECG recordings. For evaluating the classification results, we have used the ROC curve and four performance measures: Accuracy, Area Under Curve (AUC), sensitivity, and specificity.