Early Cardiac Arrest Detection Based on Open-Source Heart Sound Databases

The improvement of stethoscope technology has enabled the patient to have high-quality heart sound recordings. The heart sounds of healthy and unhealthy people are recorded using an electronic stethoscope. Sounds of heart have been deeply analyzed and demonstrated to know the importance of detecting pathologies in clinical applications. The use of cardiac auscultation systems is among the most economical methods for finding and diagnosing a variety of heart problems. The system can assist doctors in making better decisions by primarily detecting the presence of excess or aberrant cardiac waves. Unfortunately, there has been presently very little use of such methods in clinical studies. This is primarily a result of ignorance regarding the enormous datasets that are freely accessible. The open-access database may have been used for analyzing the heart sounds. Teaching Heart Auscultation, PhysioNet/CinC Challenge 2016, PASCAL Classifying Heart Sound Challenge, MHSDB (Michigan Heart Sound and Murmur Database), and other open-source databases are outlined in the paper, along with their main goals, training and testing datasets, and heart sounds with hand-corrected annotations. The open-heart audio datasets are useful to the scientific community because it can serve as a reference for other researchers who are interested in studying early cardiac arrest detection.