Algorithm submission error

Dear organizers,

I was trying to make a test submission and got the following error:

• FileNotFoundError: [Errno 2] No such file or directory: '/input/clinical-information-prostate-mri.json'

And when creating a real submission got also an error but couldn't see it. That seems to be an error naming the clinical information json, could you review it?

Hi Jesús,

We fixed that, you can try again. The name of the JSON is /input/psa-and-age.json. Sorry for the inconvenience.

Best regards,

I am wondering if we could have a dev phase, we could build and test our docker. Just a single image in the input dir is fine.

Or you could simple increase the number of submissions in the validation phase.

Also I am wondering why we have nii images but the test set contains mha images ?

It makes our docker build even harder to test.

PS: I am on the way to create an algorithm but the docker build failed even I just submitted the public template https://github.com/chaimeleon-eu/OpenChallenge. Could you kindly let me know how I could get the docker build logs ?

Hi,

Regarding your question about a dev phase, this is something that we can't add now, just for you to know, the failed submissions aren't taken into consideration on the limit of submissions that you have.

About the nii and mha images, this is something internal to Gran-Challenge, which internally converts nii to mha. In any case, it should not be a problem since using libraries like SimpleITK the read output is the same regardless of the image format.

Finally, about the docker build logs, if the submission has failed, the Grand-Challenge platform itself specifies where the error is in the logs.

Let us know if you have more questions, hope we could help.

"About the nii and mha images, this is something internal to Gran-Challenge, which internally converts nii to mha. In any case, it should not be a problem since using libraries like SimpleITK the read output is the same regardless of the image format."

I found a significant difference between my local CV and the LB. Could you kindly provide me a script that was used to convert from nii to mha? Otherwise just a pair of (nii, mha) images should be fine (I think I know why but would like to confirm the issue).

Many thanks.

Hi, we reached grand-challenge to get their support on this, since this happens on their platform and the response we received is that on the platform they use this library to convert all all images, it's freely accessible via panimg · PyPI: https://github.com/DIAGNijmegen/rse-panimg

Hope this helps!