A ResNet-Attention Approach for Detection of Congestive Heart Failure from ECG Signals

Congestive Heart Failure (CHF) is a prevalent and potentially life-threatening cardiovascular condition affecting millions worldwide. Early and accurate diagnosis of CHF is crucial for effective patient management and improved outcomes. An electrocardiograph (ECG) is a non-invasive and widely available tool for assessing cardiac function. This study presents a practical approach to extract the ECG cycles and automatically classify CHF beats using attention-based Residual Networks (ResNet). Our approach leverages the power of ResNet to extract hierarchical features from ECG signals and capture relevant patterns indicative of CHF. Furthermore, we introduce an attention mechanism that dynamically highlights informative regions within the ECG beats, allowing the model to focus on critical parts contributing to accurate classification. We conduct experiments on ECG recordings, extracting both normal and CHF beats. Our results demonstrate the superiority of the proposed ResNet-Attention model over some of the recently proposed methods, achieving a more balanced accuracy, sensitivity, and specificity of 93.25%, 92.30%, and 94.11 %, respectively, in normal and CHF beat classification. The ResNet-Attention approach presented in this study shows potential in detecting CHF beats, ultimately aiding clinicians in making informed decisions and improving diagnosis. The robustness and interpretability of our model make it a valuable tool for real-world clinical applications in cardiovascular medicine.