Deep Learning Models for Coronary Artery Disease Classification using Phonocardiograms

This research paper presents a comparative analysis of three deep learning models for automatic heart sound classification using spectrograms as features. The models investigated include Custom CNN, Transfer Learning based on pretrained CNN Architectures i.e., VGG16 and MobileNetV2, and PSO-optimized CNN. The performance of these models is evaluated in terms of precision, recall, and F1-score for both normal and abnormal heart sounds. The results are compared with previous studies that utilized different feature extraction techniques and classification algorithms. The proposed work achieves promising results i.e., 98% of F1 score, demonstrating the effectiveness of using spectrograms and Transfer Learning for heart sound classification.