COVID-19 dataset 3 classes

Data is in the form of JPG X-ray images. We are using a collection of the two datasets from the Kaggle Chest X-rays and the IEEE8020 COVID-19 Chest X-ray dataset provided by Dr. Cohen from John Hopkins Hospital.These datasets were combined for comparing the healthy patients, bacterial pneumonia patients and COVID-19 virus-induced pneumonia patients. The COVID-19 Chest X-ray data is collected by Dr. Joseph Paul Cohen of the University of Montreal. Both of these datasets consist of posterior-anterior chest images of patients with pneumonia. The COVID-19 dataset is being updated daily as more cases are published. For this study, the dataset was accessed on the 18th of March, 2020. The dataset was split into 4  different categories, with around 60-70 X-ray images per class and with 9 X-ray images per class used as a test set: Healthy, Pneumonia (Viral), Pneumonia (Bacterial) and Pneumonia (COVID-19). These two datasets collectively consist of 270 X-ray images available for training and total 36 images for testing. For our model, we will be targeting three classes, typically Healthy Normal Patients, Pneumonia Bacterial and Pneumonia caused due to COVID-19 virus traces. These X-ray images have been in non-uniform sizes.