This paper presents an initial open image dataset for COVID-19, consisting of 123 frontal view X-rays. The dataset was compiled from medical images available on websites and publications, with the aim of providing a resource for computational analysis of COVID-19 cases. The dataset includes images of patients with pneumonia, including cases of COVID-19, MERS, SARS, and ARDS. The images are collected from public sources to ensure patient confidentiality is maintained. The dataset is available at https://github.com/ieee8023/covid-chestxray-dataset. The dataset is expected to help in studying the progression of COVID-19 and how its radiological findings differ from other types of pneumonia. It can be used to develop tools for predicting the type and outcome of pneumonia, as well as for triaging cases in the absence of physical tests, particularly when PCR tests are in short supply. These tools could help predict patient survival and assist physicians in managing care. Additionally, the dataset can monitor the progression of COVID-19 in positive patients, aiding in better tracking of their condition. The dataset includes metadata describing each image, such as the patient's demographics, the type of imaging (PA, AP, etc.), and the findings. The data is largely compiled from websites such as Radiopaedia.org, the Italian Society of Medical and Interventional Radiology, and Figure1.com. Images are extracted from online publications, websites, or directly from PDFs using the tool pdfimages. The goal is to maintain image quality during this process. The dataset is derived from various papers and studies, including those published in journals such as The Lancet, New England Journal of Medicine, and Radiology. The dataset aims to provide essential data for training and testing deep learning-based systems, which could improve the identification of COVID-19 and other types of pneumonia. The dataset and its analysis could help better understand the dynamics of the disease and improve treatment strategies.This paper presents an initial open image dataset for COVID-19, consisting of 123 frontal view X-rays. The dataset was compiled from medical images available on websites and publications, with the aim of providing a resource for computational analysis of COVID-19 cases. The dataset includes images of patients with pneumonia, including cases of COVID-19, MERS, SARS, and ARDS. The images are collected from public sources to ensure patient confidentiality is maintained. The dataset is available at https://github.com/ieee8023/covid-chestxray-dataset. The dataset is expected to help in studying the progression of COVID-19 and how its radiological findings differ from other types of pneumonia. It can be used to develop tools for predicting the type and outcome of pneumonia, as well as for triaging cases in the absence of physical tests, particularly when PCR tests are in short supply. These tools could help predict patient survival and assist physicians in managing care. Additionally, the dataset can monitor the progression of COVID-19 in positive patients, aiding in better tracking of their condition. The dataset includes metadata describing each image, such as the patient's demographics, the type of imaging (PA, AP, etc.), and the findings. The data is largely compiled from websites such as Radiopaedia.org, the Italian Society of Medical and Interventional Radiology, and Figure1.com. Images are extracted from online publications, websites, or directly from PDFs using the tool pdfimages. The goal is to maintain image quality during this process. The dataset is derived from various papers and studies, including those published in journals such as The Lancet, New England Journal of Medicine, and Radiology. The dataset aims to provide essential data for training and testing deep learning-based systems, which could improve the identification of COVID-19 and other types of pneumonia. The dataset and its analysis could help better understand the dynamics of the disease and improve treatment strategies.