Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add softseg support for nnUNet #16

Open
valosekj opened this issue Jun 22, 2023 · 3 comments
Open

Add softseg support for nnUNet #16

valosekj opened this issue Jun 22, 2023 · 3 comments
Assignees
Labels

Comments

@valosekj
Copy link
Member

Originily posted in #13 (comment) by @tzebre:

About the naga comment for softseg support :

But, if we can find a way to encode the soft segmentations in a different range (not so 0-1) that can result in integer value after binarizing, yes, we can indeed use soft segs with nnuent. Hope this makes sense! (and it would really cool if we are able to do this!)

I have put the softseg segmentation support aside for the moment, it's going to take me a while. But I think we can merge without this feature for now.

@tzebre
Copy link
Member

tzebre commented Jun 22, 2023

My first thought is to discretize softseg into integer value class (number of class given in args).
I need to check the maximum amount of different voxel value.

@tzebre
Copy link
Member

tzebre commented Jun 23, 2023

I have created a new branch, the code is here

@valosekj
Copy link
Member Author

I have created a new branch, the code is here

Awesome! 👍🏻 I took the liberty to open draft PR for easier code check/review.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants