Key result
CapsNet achieves ~92% accuracy for automated heart murmur classification.
Why the study?
Convolutional neural networks may face challenges classifying the MFCC spectrum of heart sounds without good feature extraction techniques.
Does a Capsule neural network (CapsNet) improve the prediction accuracy of heart murmur classification compared to other CNNs?
Population
Heart sound recordings from the 2016 PhysioNet database and a clinical auscultation dataset
Comparison
Capsule neural network vs other convolutional neural networks
Design
Machine learning model development and validation study
Authors
Loading...
High accuracy supports exploration of CapsNet for point-of-care murmur screening; leaves open head-to-head comparisons with conventional CNNs.
Does a Capsule neural network (CapsNet) improve the prediction accuracy of heart murmur classification compared to other CNNs?
A capsule neural network (CapsNet) demonstrates high accuracy (up to 91.67%) in classifying heart murmurs from audio recordings, offering a potential improvement over traditional CNNs.
Tsai et al. (2023) studied Heart murmurs. Capsule neural network (CapsNet) vs. Other convolutional neural networks was evaluated on Prediction accuracy of heart murmur classification. A capsule neural network (CapsNet) demonstrated high feasibility for automated heart murmur classification, achieving accuracies of 90.29% on validation and 91.67% on the test dataset.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: