Title: Enhanced Cardiovascular Disease Classification Using CNN Xception Model: A Deep Learning Approach with ECG Image Augmentation
Authors: Zakaria K. D. Alkayyali, Nor Adnan Yahaya, Samy S. Abu-Naser
Volume: 10
Issue: 9
Pages: 9-17
Publication Date: 2026/09/28
Abstract:
Cardiovascular diseases (CVDs) remain a leading cause of mortality worldwide, emphasizing the need for early and accurate diagnosis. In this study, we propose an advanced deep learning approach for the classification of CVDs using the Xception Convolutional Neural Network (CNN) model, leveraging augmented ECG images. The dataset, collected from Kaggle, comprises 928 ECG images categorized into four classes: patients with a history of myocardial infarction (MI), patients with abnormal heartbeats, patients with MI, and normal ECGs. Image augmentation was employed, increasing the dataset to 4000 images. The dataset was split into 60% training, 20% validation, and 20% testing sets. Our model was trained over 80 epochs using Python, with performance evaluated using F1-score, recall, precision, and ROC-AUC metrics. The proposed model achieved an impressive F1-score of 98.00%, recall of 98.30%, and precision of 98.23%, demonstrating its effectiveness in accurately classifying ECG images. These results indicate the potential of deep learning techniques for improving diagnostic accuracy in cardiovascular disease