Enhancing Early Heart Disease Detection: Comparing CNN, RNN, and Hybrid CNN-RNN Models for Improved Diagnostic Accuracy

Abstract Early detection of heart disease is essential for timely intervention and improved patient care. This study evaluates the diagnostic accuracy of Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and a Hybrid CNN-RNN model using an ECG image dataset of cardiac patients obtained from Mendeley Data. The dataset comprises 929 ECG images, including 284 normal and 644 diseased cases. Preprocessing involved resizing and normalization of ECG images. CNN layers were utilized for spatial feature extraction, while RNN layers captured sequential patterns. The hybrid architecture integrated these strengths for enhanced diagnostic precision. Experimental results revealed that the CNN model achieved 65.232% accuracy, the RNN model 65.230%, and the Hybrid CNN-RNN model 92%, with superior sensitivity and specificity. These findings demonstrate the potential of hybrid models for advancing early detection of heart disease, paving the way for more effective computational tools in healthcare diagnostics.