Machine Learning and Deep Learning-Based Detection and Analysis of COVID-19 in Chest X-Ray Images

Machine learning (ML) is a cutting-edge method with numerous applications in prediction and classification. This technology should be used to identify high-risk patients, their death rates and other irregularities in the COVID-19 pandemic (Taresh et al. in Int J Biomed Imaging, 2021 [1]). ML can be used to learn more about the virus’s nature and to foresee potential problems. With the goal in mind to help the healthcare sector, we can definitely leverage the advancement of technology (Chowdhury et al. in IEEE Access 8:132665–132676, 2020 [2]). This paper uses the COVID-19 dataset available on Kaggle. Various machine learning techniques are used to weigh the risk of COVID-19 disease in a patient in the proposed work. VGG19, MobileNetV2, DenseNet201, CapsNet201, COVID-Net, CoroNet and VGG16 are tested for classifying the images of normal human lungs versus lungs affected by viral pneumonia due to COVID-19. The performance of various machine learning algorithms is analysed, and it was determined that VGG16 algorithm achieved the best accuracy (97%) in tests.