nnUNet icon indicating copy to clipboard operation
nnUNet copied to clipboard

Preprocessed images in nii.gz format

Open Jamshidhsp opened this issue 3 years ago • 3 comments

Hi, Thank you for your great work. I am going to compare nnUNet model with some other architecture in case of accuracy. To do this and have a fair comparison, I need to have the pre-processed images. I am aware that they could be found under the pre-processed folder. Do you think if there is a way to have them in .nii.gz format? Thank you

Jamshidhsp avatar Jul 19 '22 08:07 Jamshidhsp

Hi, Thank you so much for your work on this project, it is very amazing, I learned a lot.

I too have the same question about whether we can get preprocessed data in .nii.gz format. @Jamshidhsp Please kindly share with me if you found a method to get preprocessed data in .nii.gz format. Thank you.

LokeshaRasanjalee avatar Jul 26 '22 19:07 LokeshaRasanjalee

Hey @LokeshaRasanjalee, I tried to load npz files in numpy and then saving them in .nii.gz format using nibabel library.

Jamshidhsp avatar Jul 27 '22 07:07 Jamshidhsp

@Jamshidhsp Great. Thank you very much !!!

LokeshaRasanjalee avatar Jul 28 '22 15:07 LokeshaRasanjalee

Remember to restore the geometry (spacing etc) when exporting them ;-)

FabianIsensee avatar Aug 30 '22 10:08 FabianIsensee

@FabianIsensee Thank you for your hint. Could you please explain more a bit. I couldn't understand restoring! I am going to compare results from another architecture (apart from nnUNet) with the results achieved by nnUNet. Is it the wrong way to simply load *.npz files and save them in *.nii.gz format?

Jamshidhsp avatar Aug 30 '22 13:08 Jamshidhsp