Hello, dear organizers, thanks for your efforts for DFC2024.
We noticed that the test dataset was described as "There are 306 tiles in the training set, 125 tiles in the validation set, 126 tiles in the test set".
However, we found that the test data has been changed and the new version contains more than 400 tiles. We wonder which version is the real test dataset?
Hello,
You're right, the test dataset has been reworked, mainly by adding new independent data, to limit the correlations reported by some participants during phase 1.
It contains now 460 patches. We will update the description on Codalab, thank you for pointing out this problem.
Best regard