Closed cyh-0 closed 3 years ago
Hi @Ch1kara, thanks for your question! We used full CMMD dataset for evaluation, i.e. all 1,775 studies. This is a relatively new dataset so none of the models in the meta-repository were trained using CMMD data. Please refer to the TCIA website for CMMD dataset details/download: https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=70230508.
Thank you so much for your explanation!
Hi,
I am wondering how did you divide the train/test/validation for the CMMD dataset. Is there any file list that you can provide?
Cheers