Prediction of Cardiomegaly Disease Using Deep Learning

Cardiomegaly can be viewed as an indication of an underlying medical condition rather than being a separate ailment. Within the scope of diagnostic assessments, identifying an enlarged heart within a chest X-ray image signifies the presence of cardiomegaly. The etiology of cardiomegaly encompasses a spectrum of causative factors, including hypertension, coronary artery disease, infections, hereditary abnormalities, and cardiomyopathies. Timely identification and accurate diagnosis of cardiomegaly hold the utmost significance in facilitating efficacious treatment approaches and optimizing patient outcomes. The current study proposes a deep learning-based methodology that leverages medical imaging data to predict the occurrence of cardiomegaly. Specifically, this method employs Deep Convolutional Neural Network (DCNN) architecture to autonomously classify and anticipate the presence of cardiomegaly in chest X-ray images. The training and evaluation of our proposed model are conducted using the National Institutes of Health chest X-ray dataset, which is available on the Kaggle platform. Our system demonstrates a notable level of accuracy, achieving an average accuracy rate of 97%, a precision accuracy rate of 97%, a recall accuracy rate of 96%, and an average F1-score accuracy rate of 97% for identifying cardiomegaly in CXR images.