I am using a combination of 4 opensource COVID-19 datasets.
- https://github.com/ieee8023/covid-chestxray-dataset
- https://github.com/agchung/Figure1-COVID-chestxray-dataset
- https://github.com/agchung/Actualmed-COVID-chestxray-dataset
- https://www.kaggle.com/tawsifurrahman/covid19-radiography-database
To start, execute the following commands to clone the github repositories. Make sure your current directory is the clone of my repo.
git clone https://github.com/ieee8023/covid-chestxray-dataset
git clone https://github.com/agchung/Figure1-COVID-chestxray-dataset
git clone https://github.com/agchung/Actualmed-COVID-chestxray-dataset
The Kaggle dataset must be downloaded and unzipped to the same location. Navigate to where the Kaggle dataset was downloaded, and find the 'COVID-19' and 'NORMAL' folders. Create a 'VERIFICATION' file alongside them.
Next, run makedataset.py. It will move all one of three folders in 'chest x-rays into the COVID-19 Radiography Database'. If it was a positive test, the image is in 'COVID-19', and if it was normal, the image is in 'NORMAL'. Certain images designated as verification have been placed in verification.txt and will be moved to the 'VERIFICATION' folder for testing of our neural network. Any images in 'VERIFICATION' will not be used for training.