A CNN Model for Detecting Coronavirus in Chest Computed Tomography Scan Images using Gabor Filter in Pre-processing Stage

Abstract

Coronavirus is the cause of the pandemic illness. The Reverse Transcription–Polymerase Chain Reaction (RT-PCR) test is frequently used to identify coronavirus. On Computed Tomography (CT) images, the extent to which the virus has impacted the lungs can be seen clearly. In 15 minutes, CT data are accessible, but RT-PCR takes 24 hours. The proposed model looks for the virus in the lungs, which is more accurate than PCR, which only looks for it in the nose or throat. More accurate and dependable data can be obtained, if Computed Tomography scans are employed. The proposed innovative model has an accuracy with Gabor filter and without Gabor filter is 0.83 and 0.75 in recognizing the coronavirus in Lung Computed Tomography Scans. The accuracy of the preceding models VGG16, VGG19, ResNet50, and Mobile Net with the Gabor filter is 0.79,0.81,0.81,0.81 and 0.68,0.61,0.71 and 0.79 without it. Gabor filter is a linear filter that is sensitive to orientation and can assist reduce noise from data. Our model obtains an accuracy of 0.83, which is higher than the Gabor Filter models VGG16, VGG19, ResNet50, and Mobile Net.