AI Powered COVID-19 Detection with Optimal Feature Analysis in Flask-Based Diagnostic System

Abstract

The integration of Internet of Things (IoT) technology is crucial for advancing healthcare solutions, particularly in the context of COVID-19 detection and management. IoT systems can enhance real-time monitoring and data collection, offering timely insights and improving patient outcomes through connected devices and intelligent analysis. Existing challenges in COVID-19 detection include the need for accurate and rapid classification methods amidst large volumes of health data and the limitations of traditional diagnostic approaches. So, this research addresses these challenges by developing a comprehensive IoT-based COVID-19 detection system utilizing the Optimal Iterative COVID-19 Classification Network (OICC-Net). The procedure begins with the establishment of a flask environment to support web-based interactions for both administrators (doctors) and users (patients). The system features distinct registration and login modules for these user roles, ensuring tailored access and functionality. In the admin environment, the OICC-Net is implemented using the San Francisco COVID-19 dataset, which is subjected to rigorous preprocessing. Feature extraction is enhanced through a hybrid method combining Random Forest Infused Particle Swarm-based Black Widow Optimization (RFI-PS-BWO), which optimizes the selection of relevant features. This is followed by Iterative Deep Convolution Learning (IDCL) for further feature refinement. The Convolutional Neural Network (CNN) classification model then categorizes the data into "no virus," "other virus," and "SARS-CoV-2 (SC2) virus" classes. The trained model is subsequently saved and evaluated to ensure its performance meets the required accuracy and reliability standards. The final component involves user interaction, where patients submit their health details via the application, and the OICC-Net model provides rapid and precise predictions of COVID-19 status. This approach effectively integrates advanced machine learning techniques with IoT infrastructure, facilitating accurate and efficient COVID-19 detection and classification in real-time scenarios.