Visualization of the segmentation results codes

[Eurus-Holmes]

Would you release your visualization codes?

[cshwhale]

We visualized the segmentation results by ITK-SNAP software.

[Eurus-Holmes]

oh, I see, but in your code test.py, it seems you don't save the predicted mask images?

[pervinh]

Hi, I have the same problem too. How can i save segmentation results as nifti files?

[cshwhale]

@Eurus-Holmes
@pervinh
Hi, I save the predicted mask as nii format and then visualize it with ITK-SNAP software. The code is as follows:
import nibabel
vis_data = nibabel.Nifti1Image(np.array(mask).astype(np.float32), np.eye(4))
nibabel.save(vis_data, "mask_vis.nii.gz")

[pervinh]

@cshwhale
Thank you, this worked

[Eurus-Holmes]

@cshwhale Thanks!

[ni-wei]

@Eurus-Holmes
@pervinh
Hi, I save the predicted mask as nii format and then visualize it with ITK-SNAP software. The code is as follows:
import nibabel
vis_data = nibabel.Nifti1Image(np.array(mask).astype(np.float32), np.eye(4))
nibabel.save(vis_data, "mask_vis.nii.gz")

Hi,

in order to register the predicted mask with the data & label, does it make sense to obtain a "target_affine" from the data, and replace the "np.eye(4)" in SH's code with this "target_affine"?
I obtained the "target_affine" by inserting the following 2 lines next to https://github.com/Tencent/MedicalNet/blob/25ba32573f281646f8a3d15f096d017de615ee3a/test.py#L62
data = nib.load(os.path.join(sets.data_root, img_names[batch_id]))
target_affine = data.affine

Thanks,
Wei