Closed mayaliliya closed 2 years ago
A couple of minor points with respect to image and patient counts, but otherwise this looks pretty straightforward. I'm good to approve this once the concerns over counts are addressed.
Thanks for the replies, I'm approving this now. The only thing to maybe consider is dropping the patient count down to 16648, as strictly speaking that's our actual total patient count.
Thanks for the replies, I'm approving this now. The only thing to maybe consider is dropping the patient count down to 16648, as strictly speaking that's our actual total patient count.
Sounds good, I dropped it back down to 16,648. If there are questions regarding it moving forward we can more explicitly document the discrepancy as for now I will simply drop the total by 10.
Released the COVIDx9 dataset that maintains the same test set as COVIDx8 but has additional normal, pneumonia, and COVID-19 images into the training dataset from BIMCV, RSNA, and Stong Brook public datasets.