[Insight-users] Control of deformation field generation via landmarks?

[Luis Ibanez]

Hi Ruben,

The KernelTransform used inside this filter *already* blends
the deformation from one landmark to the next.

You may want to play with replacing the reciprocal elastic body
kernel with the elastic body one.

If you want to see profiles of the different kernels you can
use the application:

         InsightApplications/ThinPlateSplines


You should also play with the value of Stiffness. This parameter
allows to modify the behavior of the KernelTransform from
interpolation to approximation. The default behavior is to
interpolate.


When applying deformations it is wise to add landmarks on the
outside of the image, and to have their homologous landmarks
to be superimposed, so that they generates a constrain of
zero deformation around the image.



Regards,


     Luis


-----------------------
Ruben Schilling wrote:
> Hi,
> 
> I am trying to evaluate some particular elastic and large  deformations 
> using landmarks. I have assigned a modest number of  landmarks (about 
> 30) sitting in image volumes of size 400 x 400 x  250. Considering the 
> particular type of (elastic) deformation I am  facing the landmarks seem 
> to cover the deformation pretty well.
> I have built a prototype program around the example  
> LandmarkWarping2.cxx, which I have modified for 3D and additional  
> optional large (sigma ~ 50) gaussian regularization of the  deformation 
> field from the spline. I am using a reciprocal elastic  body kernel 
> (without stiffness).
> 
> 
> I am having two kinds of trouble though:
> (1) Scanning through the image slices, it seems that they are  shifting 
> from right to left (which is also the main direction of my  landmark 
> positions)
> (2) The biggest portion of the moving image is cut off (about 60 out  of 
> 250 slices remaining), primarily about 20 slices before and after  the 
> slice-block with the landmarks stay in the game
> 
> Since my landmarks do not cover the whole image depth I would have  
> expected parts not covered by landmarks to be moved via spline  
> regularization estimated from the sparse deformation field of the 30  
> landmarks.
> 
> 
> Is there a way to increase the radius of regularization (e.g. some  
> linear weigth of distance to next landmark) or any control of the  
> regularization at all?
> My guess is, that my gaussian regularization looses too much strength  
> over distance. Is there another (maybe direct?) way to control the  
> deformation field interpolation with the kernel transforms?
> 
> I would be grateful for advice on this matter.
> 
> Best,
> Ruben
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>