Warning message when loading my segmentation NIfTI (.nii) file (ITK-SNAP 3.2.0)

[Atilla Gül]

Hi,

Currently, I'm using the ITK-SNAP software with version number 3.2.0. After uploading my CBCT DICOM-image series, I get the following warning message when I load my segmentation as a NIfTI (.nii) file: 'There is a mismatch between the header of the image that you are loading and the header of the main image currently open in ITK-SNAP. The images have different spacing. ITK-SNAP will ignore the header in the image you are loading'. 

How could I prevent this problem and if this is not possible, could this affect the accuracy in the segmented data?

Kind regards,

Atilla Gül

[Philip A Cook]

This is probably caused by a small error in the spacing stored in the Nifti header. Sometimes I save something with spacing of say, 0.8 mm, and it is stored in the header as 0.79999 mm. You can verify this by looking at the spacing of both images, if you load the segmentation as a gray image you can see its header info under the info / metadata tab on the layer inspector.

It shouldn't affect your segmentation within SNAP as SNAP will ignore such differences, and rely instead on the gray image header to interpret the segmentation image. It becomes an issue if you do anything that interprets the segmentation image as a point set - but then the error should be very small.

I'm not sure if there's a way to avoid the error entirely. Saving your grayscale image as Nifti might do the trick as then both headers will hopefully be identical again.

> --
> You received this message because you are subscribed to the Google Groups "itksnap-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to itksnap-user...@googlegroups.com.
> To post to this group, send email to itksna...@googlegroups.com.
> Visit this group at http://groups.google.com/group/itksnap-users.
> For more options, visit https://groups.google.com/d/optout.

[Sandhitsu Das]

You can try to set the header of the segmentation to match the gray image using "c3d gray.nii seg.nii -copy-transform -o seg_header_matched_to_gray.nii"

Sandy