AI-Enhanced ECG Intelligence: Optimized Deep Learning Pipelines for Early and Reliable Detection of Cardiac Disorders

The early and accurate identification of heart diseases is a fundamental requirement of smart healthcare systems, where artificial intelligence (AI) can significantly enhance diagnostic reliability and clinical decision support. This study proposes an AI-based electrocardiogram (ECG) processing methodology that facilitates effective signal analysis and improves the early diagnosis of atrial fibrillation (AFIB) and myocardial infarction (MI). Traditional machine learning approaches, including Support Vector Machines, Naïve Bayes, and Decision Trees, are evaluated alongside deep learning models, with particular emphasis on a customized convolutional neural network (CNN) optimized for robustness and generalization. Experiments conducted on the MIT-BIH, PTB-DB, and SVDB datasets demonstrate that the proposed CNN consistently outperforms traditional models, achieving an average accuracy and F1-score of 99% for arrhythmia classification. For combined AFIB and MI evaluation, the CNN attains F1-scores of 99% and 97%, respectively. These results highlight the potential of deep learning-based ECG intelligence for smart healthcare systems, while future work will address segmentation and peak-detection challenges to further enhance real-time diagnostic reliability.