[Insight-users] Demons deformable Registration: initialization and RoI

[Karthik Krishnan]

Hi Jan,

1. I don't think its worth representing an affine or similarity 
transform with BSplines. I don't think this is even possible. You might 
just use the transformed image from the first stage as the initial image 
for new registration as in (c). Maybe I am wrong here.

2. Yes it is possible to set the region of interest if you are using 
BSplineDeformableTransform

You could use the methods:
ImageRegistrationMethod->SetFixedImageRegion()
metric->SetFixedImageMask(), SetFixedImageRegion, SetMovingImageMask()

Effectively the metric iterator iterates over each pixel in the Fixed 
image that is both the FixedImageMask and the FixedImageRegion specified 
and that also after transformation maps into the mask specified through 
the SetMovingImageMask().

I don't see how you could do it with Demons other than your suggestion 2(b). 

Thanks
regards
karthik





Jan Schreiber wrote:

>Hi all, 
>
>I have two questions on demons deformable registration:
>
>1. Initialization from non-deformable pre-registration
>Before performing the deformable registration, I am using a similarity transform to pre-align the images. Now I want to use the result from the pre-registration with my deformable registration. 
>a) is there a simple way to do this?
>b) does it make sense to compute a deformation field fom the similatity transform and use this as initial deformation field, is this supported by itk
>c) should I just resample the image with the similarity transform and use the "new" image?
>
>2. Setting a region of interest for the deformable registration
>a) Is it possible to set a region of Interest to the deformable registration to not register the whole image?
>b) If not and I have to extract a region and create a new image from that - is it (easily) possible to re-integrate that region to the original (large) image?
>
>Thanks in advance, 
>Jan
>
>
>_______________________________________________
>Insight-users mailing list
>Insight-users at itk.org
>http://www.itk.org/mailman/listinfo/insight-users
>
>  
>