Screening of Cardiomegaly using Ensemble Model of InceptionV3 and MobileNet

Cardiomegaly is an asymptomatic condition characterized by an enlarged heart. It can be an indication of a possible underlying condition which could be temporary or permanent. However, some patients might show symptoms like shortness of breath, palpitations, irregular rhythm of heartbeat, swelling in belly or legs, etc. Cardiomegaly might lead to several complications such as heart failure, blood clots, leaky heart valve, or even cardiac arrest and sudden death. Early diagnosis of cardiomegaly is important for improving outcomes and reducing the risk of complications. The condition is typically diagnosed through imaging tests such as chest X-rays which can be time consuming and require expert interpretation. This article presents a machine learning model and its application in the field of biomedical engineering. It explores the use of an ensemble model that integrates the InceptionV3 and MobileNet models for this purpose and relies on the transfer learning technique. The model was able to screen cardiomegaly from posteroanterior chest x-ray images with an accuracy of 83.4%, precision of 83.5% and sensitivity of 84.7%.