COVID-Net
COVID-Net copied to clipboard
lindawangg
# Annotation samples in training and testing
Thank you for your valuable datasets, When we use the code to download the dateset, I faced the following issues: 1- I get 1579 images in the test set, but the "test_COVIDx3" has just 300 labeled images 2 I get 13696 images but the train set but the "train_COVIDx3" has just 13670 labeled images
Can you provide us with the updated text files which containing labels With best regards
Please refer to this FAQ for 1. For 2, it's because https://github.com/ieee8023/covid-chestxray-dataset dataset is constantly adding new images. train_COVIDx3.txt is what we used during training.
Thank you for response. Unfortunately, the "test_COVIDx3" still not updated, it has just 300 labeled images, while the actual images are 1579. can you pleas guide me how to find the the updated file. With best regards
test_COVIDx3.txt is correct. We used 100 for each for a balanced test set. We kept additional pneumonia and normal images in case we decide to increase the test set in the future when more covid test sample are added.
hye Im using this format of files please review and let me know plz
[image: image.png]
On Mon, Jun 22, 2020 at 4:16 AM Linda Wang [email protected] wrote:
test_COVIDx3.txt is correct. We used 100 for each for a balanced test set. We kept additional pneumonia and normal images in case we decide to increase the test set in the future when more covid test sample are added.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/lindawangg/COVID-Net/issues/78#issuecomment-647194164, or unsubscribe https://github.com/notifications/unsubscribe-auth/APZGFQQUGESVOK2S4TUW4PDRX2IEZANCNFSM4NZFSEIQ .
-- Thanks Sania