An Intelligent Systems for Heart Sound Signal Analysis: A Survey

Heart failure is the leading factor of mortality worldwide. Cardiac auscultation is one of oldest diagnostic methods to explore the condition of the heart. Mechanical activities of heart initiated by biopotentials generated by the Sino-Atrium (SA) node of the heart causes non-stationery and non-deterministic signals commonly known as Heart Sound Signal (HSS) and the visual illustration of HSS is termed as Phonocardiogram (PCG) while recorded by an electronic stethoscope. For assessing a variety of cardiac anomalies, such as anatomical defects of the heart and valvular cardiac disease, congestive heart failure, and heart rate and rhythm, PCG can offer the most useful diagnostic data. The S1 and S2 heart sounds, which are the two loudest heart sounds, often make up PCG. In both diseased and non-pathological circumstances, the PCG signal may contain murmurs, the S3 and S4 heart sounds as well as other unusual noises including clicks, snaps, and rubs.Furthermore, in past few years, it was effectively utilized for heart sound analysis. The current study is to attempt a thorough brief to summarize articles on heart sound evaluation with deep learning released in 2017–2023, since the majority of review works on the subject were completed prior to 2017. In addition to providing insights into the developments and potential avenues for upcoming research on deep learning for the analysis of cardiac sounds, this study contrasts deep learning with conventional machine learning.