A comprehensive dataset of heart sound recordings from patients with aortic, mitral, and tricuspid regurgitation, as well as healthy controls, was collected to support automated algorithm development.
This dataset provides high-quality, labeled heart sound recordings of various regurgitation types to serve as a benchmark for machine learning research and clinical auscultation training.
Heart regurgitation is a cardiac condition characterized by the backward flow of blood, producing audible murmur sounds detectable during auscultation. If left untreated, it can lead to serious complications affecting cardiac function. This article presents a comprehensive dataset of heart sound recordings, including aortic regurgitation (AR), mitral regurgitation (MR), tricuspid regurgitation (TR), and healthy heart sounds, collected from patients at a single hospital using an electronic stethoscope. For each participant, recordings were obtained from three standard chest locations, and all diagnoses were confirmed by an experienced cardiologist. The dataset provides high-quality, labeled recordings that capture the variability of regurgitation sounds across different types and locations. It is intended to support the development and evaluation of automated algorithms for detecting cardiac abnormalities, including machine learning and signal processing approaches. Additionally, this dataset offers an educational resource for medical students and trainee clinicians to practice auscultation skills, recognize different types of regurgitation murmurs, and improve diagnostic proficiency. By making these recordings publicly available, the dataset can serve as a benchmark resource for both research and clinical training in cardiac auscultation.
Fraiwan et al. (Tue,) conducted a other in Heart valve disorders (aortic, mitral, tricuspid regurgitation). Heart sound recordings dataset was evaluated. A comprehensive dataset of heart sound recordings from patients with aortic, mitral, and tricuspid regurgitation, as well as healthy controls, was collected to support automated algorithm development.