Predicting and diagnosis of COVID-19 based on IoT and machine learning algorithm

Abstract

Coronavirus Disease-2019 (COVID-19) has emerged as one of the largest and most impactful global pandemics, causing numerous negative impacts. The profound repercussions on the living conditions of individuals from 2020 to 2023 are particularly striking. Although its impact continues, the measures taken have partially mitigated its impact. The strength of the disease is closely linked to the delay in quarantining infectious individuals, allowing them to remain infected in the community. Therefore, rapid and accurate detection of COVID-19 is crucial for effective treatment and prevention. Manual identification of Chest X-ray (CXR) and Computed Tomography (CT) images is time-consuming and prone to inaccuracies attributable to human factors. Machine learning, especially its sub-branch DL, enables rapid classification of CXR and CT images related to COVID-19. Furthermore, leveraging various Internet of Things (IoT)-based approaches facilitate the rapid transmission of these images to diagnostic centers, enabling rapid detection of disease. This paper presents a framework for transmitting CXR and CT image files acquired in hospitals through IoT devices connected to the internet network to a server running DL models. The images are then classified as normal or infected. Eight pre-trained transfer learning algorithms were used in the study, with the Xception and NasNet models performing the best, achieving 89.5% accuracy. It is evaluated that the study provides promising results for COVID-19 detection in CXR and CT images using pre-trained DL models. In addition, in line with the results obtained, it was evaluated that it can be preferred for analyzing the data that may come from wearable technologies and Internet of Medical Things (IoMT) sensors.