MURMUR Detection

Heart murmur detection research focuses on automating the identification of abnormal heart sounds, improving diagnostic accuracy and efficiency. Current efforts leverage deep learning, employing convolutional neural networks (CNNs), recurrent neural networks (RNNs), and U-Net architectures, often combined with techniques like wavelet transforms and Mel spectrograms to extract meaningful features from phonocardiograms (PCGs) or other heart sound recordings. These advancements aim to reduce reliance on expert auscultation, improve accessibility to cardiac diagnostics, and potentially enable earlier detection of cardiovascular diseases. The field is also exploring methods to improve model interpretability and uncertainty quantification for enhanced clinical decision-making.