Deep Learning for ECG Classification Using Convolutional Neural Networks and VGG16

The heart is a crucial organ responsible for circulating blood through the human body's circulatory system. It is susceptible to a variety of conditions collectively known as cardiovascular diseases. Fortunately, electrocardiogram (ECG) signals generated by heartbeats can be used to analyze cardiac activity and determine whether it is normal or abnormal. In this study, we employed the VGG16 model for transfer learning and used convolutional neural networks to classify six heartbeat micro-classes from the MIT-BIH Arrhythmia database. This approach combines the strengths of both algorithms. The experimental results show that the proposed model achieves improved performance in terms of accuracy, sensitivity, F1-score, AUCROC, precision, and loss. Its accurate classification can effectively conserve medical resources, positively impacting clinical practice.