Heart Sound Segmentation with Deep Learning

Developed a deep learning approach during my internship at the Aid-Lab, University of Sydney, to locate the precise start and end points of S1 and S2 heart sounds within phonocardiogram (PCG) signals. The project focused on training a convolutional neural network (CNN) to identify these key transition points directly, moving away from more complex segmentation methods. This approach was evaluated on a large public PCG dataset to assess its performance in automating a crucial step in cardiac signal analysis.