ISIC 2018 Data Split

[salauddintapu]

hello there,
I am working on the ISIC 2018 dataset and came across your paper. But unfortunately found no information on data split. How you used the dataset for training you model? Did you used the official train-test set or you randomly splited train dataset into train and test sets?

[srperera]

Hey, for all the ISIC datasets we used the official test train splits for all the experiments to keep things standardized with the rest of the research community. However for the PH2 dataset, we did a K-fold approach just like everyone else.