Split

The spilt of dataset can be different depend on the specific senario:

Take MICCAI AMOS 2022 as an example for reference:
The training data contains complete image (nifti format) and their ground truths for the selected data sets (i.e. patients).

For the 500 CT scans of task 1, it is planned to make a 60%/10%/30% split to get 300 training cases, 50 validation cases, and 150 testing cases.

As for task 2, the 500 CT and 100 MRI scans are planed to split to get 300 CT + 60 MRI training cases, 50 CT + 10 MRI validation cases, and 150 CT + 30 MRI testing cases. The split of CT data will remain the same in both tasks.

Download Links

We provide several data sources for downloading, after emailing to us, you can get the corresponding links.