AI-Enabled Deep Learning Model for COVID-19 Identification Leveraging Internet of Things

Abstract

Addressing the COVID-19 epidemic since December 2019 has emphasized the criticality of timely and error-free identification of infected COVID-19 individuals in medical settings. To effectively combat the epidemic, the utilization of deep TL-enabled automated COVID-19 identification on CXRs is paramount. This study recommended a real-time IoT system employing ensemble deep TL to enable early identification of infected COVID-19 individuals. The system allows for real-time transmission and identification of COVID-19 suspicious individuals. The suggested IoT model incorporates several DL models, including InceptionResNetV2, VGG16, ResNet152V2, and DenseNet201. These models, stored on a cloud server, are utilized in conjunction with medical sensors to gather chest X-ray data and detect infections. A chest X-ray dataset is used to compare the deep ensemble model against six transfer learning algorithms. The comparative investigation demonstrates that the suggested approach facilitates swift and effective diagnosis of COVID-19 suspicious patients, providing valuable support to radiologists. This work highlights the significance of leveraging deep transfer learning and IoT in achieving early identification of suspected COVID-19 patients. The proposed system, incorporating a deep ensemble model, offers a practical solution for assisting radiologists in efficiently diagnosing COVID-19 cases