Multi-component options unavailable in ITK-snap 4.X

116 views
Skip to first unread message

Carl Kulseng

unread,
Aug 8, 2023, 8:50:53 AM8/8/23
to itksnap-users
Hi i have noticed that I am no longer able to average the intensities in a 4D dynamic series, it is possible in 3.8.

Attached screenshots of both 3.8 and 4.0
Untitled4.png
Untitled3.png

Jilei Hao

unread,
Aug 8, 2023, 10:05:41 AM8/8/23
to itksna...@googlegroups.com
Hi Carl,

It is by design in 4.0. 

In 3.8, 4D images were read as multi-component. That's why the 4th dimension can be averaged. But from 4.0, they are no longer read as multi-component images but as a single component 4D image. And itk-snap currently can only perform pixel-wise average on multi-component images. 

Best,
Jilei

On Tue, Aug 8, 2023 at 8:52 AM Carl Kulseng <carlk...@gmail.com> wrote:
Hi i have noticed that I am no longer able to average the intensities in a 4D dynamic series, it is possible in 3.8.

Attached screenshots of both 3.8 and 4.0

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/itksnap-users/48cdd417-4bc9-4a1d-a5bb-bf56c0f908f4n%40googlegroups.com.

Carl

unread,
Aug 8, 2023, 1:00:26 PM8/8/23
to itksna...@googlegroups.com
Hi ok thank for the clarification. Actually, being able to average all the dynamic series together made the manual segmentation more effective in our case. So we will have to use 3.8 for that feature?

You received this message because you are subscribed to a topic in the Google Groups "itksnap-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/itksnap-users/-ByDAt2K8b8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to itksnap-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/itksnap-users/CABsSX%2B%3Dkt_KAeZUaDA-RbN8urqcQrHhDypRv%2BAWU41V82n%2Bi3g%40mail.gmail.com.

Paul Yushkevich

unread,
Aug 8, 2023, 2:24:44 PM8/8/23
to itksna...@googlegroups.com
Hi Carl,

You bring up a good point that others have brought up as well. We should add in the next release an option when loading a 4D image to treat it as a 3D multicomponent image as opposed to a 3D+time image. A work around for now would be to use a NIFTI library to reformat the image as a multi-component one. If I recall correctly, a 3D+time NIFTI with 256^3 voxels and 10 time points has dimensions [256,256,256,10,1] while a multi-component NIFTI with 10 components has dimensions [256,256,256,1,10], in which case it would be a short Python script to modify the image. Also can be done using c3d, 

I think this code

c3d -mcs your_image.nii.gz -omc your_image_mc.nii.gz

should load a 3D+time image and save as a 3D multi component. If not, then try:

c4d your_image.nii.gz -slice w 1:-1 -omc your_image_mc.nii.gz

Please let me know if that helps,

Paul

Jilei Hao

unread,
Aug 8, 2023, 2:54:42 PM8/8/23
to itksna...@googlegroups.com
That's right.

On Tue, Aug 8, 2023 at 1:00 PM Carl <carlk...@gmail.com> wrote:

Carl

unread,
Aug 8, 2023, 3:32:35 PM8/8/23
to itksna...@googlegroups.com
Hi Paul, hope you are doing well. The approach you mention works fine as a work-around. 

Thank you

Reply all
Reply to author
Forward
0 new messages