Lung Infection Detection for an Internet of Medical Things Application Using Chest CT Images Through a Deep Ensemble Convolutional Neural Network Model for Smart Healthcare

Lung infections, including pneumonia and tuberculosis, are major health concerns worldwide, and, hence, the diagnosis and detection of these diseases are vital components for management and treatment. Lung infection may cause the diseases related to this particular organ, including fever, tumor, tuberculosis, and cancer. The present work presents an internet of medical things framework for detection of lung infection from chest computer tomography (CT) images, aiming to provide timely and accessible healthcare solutions. Internet of things devices capture chest CT images from patients and securely upload it to a cloud-hosted pre-trained ensemble convolutional neural network (CNN) model to detect the presence of lung infections. The employed model has been trained with a dataset that contains 2,482 images of COVID and non-COVID type, collected for post-COVID cases. This work focuses on utilizing an ensemble-based deep CNN architecture, for lung infection detection. Results are verified on a web application interface, allowing healthcare professionals to visualize and interpret the findings. The proposed model has an accuracy of approximately 98.99% and achieves improved performance than several recent models.