COVID19, Pneumonia and Normal Chest X-ray PA Dataset

The dataset is organized into 3 folders (covid, pneumonia , normal) and metadata.csv which contain chest X-ray posteroanterior (PA) images. X-ray samples of COVID-19 were retrieved from different sources for the unavailability of a large specific dataset. First, 613 X-ray images of COVID-19 cases were collected from the following websites: GitHub [1,2], Radiopaedia [3], The Cancer Imaging Archive (TCIA) [4], and the Italian Society of Radiology (SIRM) [5]. Then, instead of data being independently augmented, a dataset containing 912 already augmented images was collected from Mendeley [6]. Finally, 1525 images of pneumonia cases and 1525 X-ray images of normal cases were collected from the Kaggle repository [7] and NIH dataset [8]. A total of 4575 samples were used in the experiment, where 1525 samples were used for each case. References: [1] http://arxiv.org/abs/2003.11597 [2] https://github.com/agchung [3] https://radiopaedia.org/ [4] https://www.cancerimagingarchive.net/ [5] https://www.sirm.org/en/category/articles/covid-19-database/ [6] https://data.mendeley.com/datasets/2fxz4px6d8/4 [7] https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia [8] https://www.kaggle.com/nih-chest-xrays/data