Hybrid CNN–GRU Framework For Automated Diagnosis Of Heart Failure Using Medical Imaging

Heart failure (HF) remains one of the leading causes of morbidity and mortality worldwide, necessitating early and accurate diagnostic solutions to support clinical decision-making. Recent advances in medical imaging and deep learning have shown significant promise; however, existing approaches often struggle to effectively capture both spatial and temporal characteristics inherent in cardiac data. This paper proposes a hybrid Convolutional Neural Network–Gated Recurrent Unit (CNN–GRU) framework for the automated diagnosis of heart failure using medical imaging modalities. The CNN component is employed to extract high-level spatial features from cardiac images, while the GRU network models temporal dependencies and sequential patterns associated with cardiac functional variations. The integrated architecture enables robust feature representation, improved learning efficiency, and reduced computational complexity compared to conventional recurrent models. Extensive experiments conducted on benchmark cardiac imaging datasets demonstrate that the proposed CNN–GRU framework achieves superior diagnostic performance, attaining an accuracy of 98.2%, sensitivity of 97.6%, specificity of 98.8%, and F1-score of 97.9%, outperforming standalone CNN, LSTM-based, and traditional machine-learning methods. The results highlight the effectiveness of the proposed hybrid framework in delivering reliable, automated heart failure diagnosis, offering a scalable and clinically applicable solution for intelligent cardiovascular healthcare systems.