Automated Detection of COVID-19 with X-ray images by Neural Network Based Algorithms

In this thesis, we analyze and perform image classification on lung X-Ray images with three state of the art convolutional neural networks. The design of Inception Resnetv2, Weakly Supervised Data Augmentation, and Discriminative Filter Bank convolutional neural networks are analyzed. We conduct image classification using the aforementioned methods with clinical x-ray chest images and review the results. The image set consists of three types of lung conditions: Normal, COVID 19, and Viral Pneumonia. It is shown that these methods effectively detect differences between COVID 19 and Viral Pneumonia