COVID-Pneumonia Detection: Expert-Selected X-ray Dataset (COVIDx-CXR)

This dataset is a selection of the most relevant X-rays from the COVIDx-CXR version 8 database, representing three classes: COVID, PNEUMONIA, and NORMAL. The database contains 3,225 frontal chest radiographs with a balanced distribution of the three classes. The data was divided into training and test subsets, with 2,884 (89%) radiographs assigned to the training phase and 341 (11%) to the test phase.The chest radiographs were selected, and the annotations were verified by a radiologist to improve the selection criteria and database quality. The radiographs represent all stages of the diseases, from moderate to critical lesions, and include patients of different ages. This diversity was included to create a comprehensive database that accurately represents the general population and enhances the model's capability for generalization.COVIDx-CXR version 8 is one of the largest public databases developed as part of COVID-Net, a global open-source initiative dedicated to accelerating advances in machine learning to assist frontline healthcare workers in the fight against the pandemic Link : https://www.kaggle.com/datasets/andyczhao/covidx-cxr2?select=train.txtIt is constructed by combining various datasets from different sources :https://bimcv.cipf.es/bimcv-projects/bimcv-covid19/#1590858128006-9e640421-6711https://github.com/ml-workgroup/covid-19-image-repository/tree/master/pnghttps://sirm.org/category/senza-categoria/covid-19/https://eurorad.orghttps://github.com/ieee8023/covid-chestxray-datasethttps://figshare.com/articles/COVID-19_Chest_X-Ray_Image_Repository/12580328https://github.com/armiro/COVID-CXNethttps://www.kaggle.com/c/rsna-pneumonia-detection-challenge/datahttps://www.kaggle.com/paultimothymooney/chest-xray-pneumonia