Joseph Paul Cohen

We don't have the variants annotated. You could use the dates to try to see if there is a phenotype change but all the images come from the start if...

The license is discussed here: https://github.com/ieee8023/covid-chestxray-dataset#license Also, please read this paper: http://arxiv.org/abs/2006.11988

Just don't use the pediatric X-ray pneumonia dataset as normal examples. The model will likely confound age to predict and not be useful.

Start by reading this paper: https://arxiv.org/abs/2006.11988 Then go over this notebook to load data: https://colab.research.google.com/drive/1A-gIZ6Xp-eh2b4CGS6BHH7-OgZtyjeP2

There should be many youtube videos about how to use it.

I don't understand what the issue is. Can you post a screenshot or some logs? Try to use the Transmission client https://transmissionbt.com/

The mapping is verified with this script: https://github.com/ieee8023/covid-chestxray-dataset/blob/master/tests/test_dataloader.py So every entry with a finding that is not "todo" and has a modality of "X-ray" will have an image associated with...

> Is all data from DataLoader resized to the same size? Yes. > Did you use some feature extraction before training neural network? In all the papers we should have...

Can you make that link public? Or maybe just upload the notebook to this issue page?

Cool! Looks good! I had just released the torchxrayvision code! I have been doing something similar but I wasn't able to get good prediction performance. Can you include an AUC...