An Extended Deep Learning based Solution for Screening COVID-19 CT-Scans

Due to immense pressure on the medical sector, there is a huge chance of human error in diagnosing the report of COVID-19 patients. For the detection of COVID-19, many Artificial Intelligence-based methodologies have been proposed. This work offers an ideal approach by a fusion of deep learning classifiers and medical images to provide a speedy and accurate identification of the COVID-19 virus by analyzing the user's CT-Scan images of the lungs. In this work, an adjusted rendition of ResNet101V2 that is improved by a Feature Pyramid Network for classifying the CT scan images into the categories: normal or COVID-19 positive, has been implemented. An accuracy of marginally significant has been obtained by ResNet101V2 with Feature Pyramid Network. The proposed FPN, using the deep learning technique, assured a satisfactory performance in terms of detecting COVID-19 in CT scan and gives a testing accuracy of 97.79%. The predicted results on test-cases are highly accurate using FPN and can assist the medical professionals in making a proper judgment.