Covid detection in CT and X-Ray images using Ensemble Learning

Abstract

Coronavirus disease is an infectious disease therefore its early detection and diagnosis is important. Currently used methods take a long time and are not precise. Computed tomography (CT) and X-rays can be used to detect coronavirus. Manual evaluation of CT scans and X-rays requires expert knowledge and are time consuming. In this paper, the authors have proposed a max voting-based ensemble learning approach for Covid detection. The state of art models used in ensemble learning are InceptionNet, ResNet and EfficientNet. The combined dataset of CT scan and X-ray is used for training and testing. The proposed approach achieved sensitivity = 98.18%, specificity = 96.6%, accuracy = 97.47% and area under curve = 95.36%. These achieved results signifies the utility of the proposed methodology.