An automated coronary artery disease identification using photoplethysmography signals with deep feature representations.

This study explores deep feature representations from photoplethysmography (PPG) signals for coronary artery disease (CAD) identification in 80 participants (40 with CAD). Finger PPG signals were processed using multilayer perceptron (MLP) and convolutional neural network (CNN) autoencoders, with performance assessed via 5-fold cross-validation. The CNN autoencoder model achieved the best results (recall 96.67%, precision 96.71%, accuracy 96.11%), outperforming MLP features and time-series imaging methods (<90%). These findings highlight the efficacy of CNN-extracted PPG features, offering a low-cost, minimally pre-processed, and portable approach for CAD diagnosis confirmed by cardiac catheterization.