Efficient Anomaly Detection Algorithm for Heart Sound Signal

Cardiovascular disease (CVD) remains the leading cause of death globally, accounting for approximately 17.9 million deaths each year, according to the World Health Organization. Early detection and intervention are critical for reducing this burden. Heart sound signals, or phonocardiograms (PCGs), offer a non-invasive and cost-effective approach to assess cardiac health. This study presents a deep neural network (DNN)-based method for detecting anomalies in heart sounds. The model effectively addresses challenges such as imbalanced class distributions, overlapping features, and subtle differences between systolic and diastolic murmurs. Evaluated on the PhysioNet/CinC 2016 dataset, the proposed approach achieved a classification accuracy of 99%, sensitivity of 98.9%, and specificity of 98.5%. These results highlight the potential of DNN-based analysis as a robust tool for early screening and diagnosis of cardiovascular abnormalities..