search

MIDA-group / CoMIR

Framework for joint representation learning, evaluation through multimodal registration and comparison with image translation based approaches
Other
74 stars 11 forks source link

Questions about the training datasets #50

Closed lajipeng closed 3 years ago

lajipeng commented 3 years ago

Hello! Thanks for your great contributions! However, it seems that there is only evaluation datasets. E.g. how can we get the trainning datasets of Zurich?

lajipeng commented 3 years ago

I noticed that your published datasets were divided into 3 sub-groups, which can enable 3-folded cross-validation, but it was still unclear to me that how to get the training datasets. I would appreciate it if you could help me !

npielawski commented 3 years ago

Hello Alan!

Thanks for your interest!

The link is indeed hard to find; it is in the first cell of the notebook: https://github.com/MIDA-group/CoMIR/blob/master/train-zurich.ipynb

The link to the Zurich dataset is: https://drive.google.com/file/d/0B9xP9Y5JKJz0ay1HdG4tRm1mME0/view?usp=sharing Feel free to use our dataloader in the 5th and 6th cells of the notebook above.

Note: you do not need to use the ground_truth data as it contains segmentation masks, unless you want to work with masks in your research. Your call 😉

Feel free to reopen if you have an issue.