hi, question about the image size of .nrrd format.

[Jihunlee326]

Thanks for your github.

I'm testing with .nrrd format. however, there is an issue where the resulting image is enlarged. (ex: input size:256x256x256 --> result : 128x128x128) Also, the image is enlarged to 32-bit or greater precision. What part of main.py should I edit?

[MontaEllis]

The size of the result will strictly correspond to the input, so please check your input and output carefully.

[Jihunlee326]

My input shape is 256x256x256. and I set hparam.py like this:

crop_or_pad_size = 256,256,256
patch_size = 128,128,128

fold_arch = '*.nrrd'

save_arch = '.nii.gz'

source_train_dir = '../datasets/source_dataset'
label_train_dir = '../datasets/label_dataset'
source_test_dir = '../datasets/source_dataset_test'
label_test_dir = '../datasets/label_dataset_test'

Is there something I'm missing out on? Please advise.

[MontaEllis]

The setting is correct, the results of the network would be 128128128 and the results after splicing would be 256256256.