[slicer-users] Comparing difference between two scans

Dominik Meier meier at bwh.harvard.edu
Fri Apr 16 14:57:11 EDT 2010

There's several metrics you can try; depends on whether you expect  
the differences to be global in nature or spatially variable.
I would start by co-registering the two images with the automated  
registration methods and then look at the differences and integrate  
them spatially to some metric.
- If you expect zero distortion differences you can limit the  
registration to 'rigid' (6DOF)
- If you run an affine registration, you can use the scale and shear  
parameters that result as an estimate of the global distortion  
- there's a "Subtract Images" function in the Filtering/Arithmetic  
menu that you can use to  get a difference image. You may have to  
normalize intensities first for this.
- If you expand to non-rigid registration (BSpline module) you then  
look at the deformation field that is being produced as a spatial map  
of the difference.
- if your final goal is an outcome marker such as a structure volume  
you can bypass registration by segmenting both scans and comparing  
- there's also the 'ChangeTracker'  module that provides different  
metrics based on deformation or intensity



Brigham & Women's Hospital / Harvard Medical School
221 Longwood Avenue, RF396, Boston, MA 02115

On Apr 16, 2010, at 2:30 PM, Murat Maga wrote:

> Hi,
> This is not really a Slicer question, but I thought I might get some
> pointers towards the right direction.
> We would like to compare scans from two different microCT scanners.
> These are not medical scanners and are not calibrated to any standard.
> We want to measure how much difference there is at the end by using  
> the
> standard procedures of the two facilities. For that, we thought of
> looking at the trabecular bone structural parameter estimates, and
> measure the extend they differ (if they do at all).
> But given that we will be scanning identical specimen, I thought there
> might be a better measure of overall difference. In theory, these two
> datasets should register perfectly and any difference can be  
> attributed
> to the difference in scan quality and/or reconstruction parameters.  
> But
> I don't know how we can convert this difference in a some kind of  
> metric.
> Any suggestions?
> M
> -- 
> A. Murat Maga, PhD
> Senior Fellow
> UW Pediatrics
> Division of Craniofacial Medicine
> 1959 NE Pacific St RR234
> Seattle, WA 98195
> 206-616 9703
> _______________________________________________
> slicer-users mailing list
> slicer-users at bwh.harvard.edu
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to slicer-users- 
> request at massmail.spl.harvard.edu with unsubscribe as the subject

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://massmail.spl.harvard.edu/pipermail/slicer-users/attachments/20100416/bcdc99ef/attachment.html>

More information about the slicer-users mailing list