Multi-Disease Cardiovascular Detection from ECG Signals Using an Attention-Driven Deep Network

Electrocardiography (ECG) is widely used to diagnose cardiovascular diseases (CVDs), particularly during prescreening. We propose a novel deep learning architecture for classifying multiple CVDs by integrating convolutional layers, residual networks, and attention mechanisms into a unified model. The motivation is to enhance the performance of traditional diagnostic methodologies by employing convolutional neural networks (CNNs) with residual connections to mitigate the vanishing gradient problem, allowing the model to learn complex patterns within the ECG signals while reducing the risk of overfitting. The proposed model integrates attention mechanisms to identify the most relevant features within ECG signals for classification. This model effectively captures both local and global features within ECG data, facilitating a comprehensive analysis of intricate cardiac patterns. Our extensive experimental results demonstrate that the proposed model effectively achieves an average classification accuracy of 99.54%, which is superior to existing deep learning-based models, and enables the detection of multiple heart conditions from a single ECG reading.