Hybrid Ensemble Learning Model for Precise COVID-19 and Pneumonia Detection with CT Scans

Abstract

Coronavirus disease (COVID-19) or C-19 is caused by the SARS-CoV-2 virus. Most people infected with the virus will experience mild to moderate respiratory symptoms and will recover without requiring specific treatment. COVID-19 has increased the need for accurate diagnosis, prompting researchers to create more advanced and efficient detection technologies. Currently, many investigations are being conducted, including reverse transcription PCR tests, chest radiographs, ultrasound scans, and CT scans. They are best conducted later in the illness phase when sensitivity and specificity are max. In this work, the adaptive normalization and enhancement (ANE) technique is proposed for pre-processing. It normalizes pixel intensity values, enhances contrast, and reduces variability in image quality. Deep convolutional feature mapping (DCFM) is employed to automatically learn and extract comprehensive features in pre-processed CT scans. Finally, hybrid ensemble learning model (HELM) is proposed to increase the accuracy and reliability of COVID-19 and Pneumonia identification, resulting in better patient outcomes and more effective pandemic management.