[Insight-users] Registration of two anisotropic volumes

[Karthik Krishnan]

Yan Yang wrote:

> Dear All,
>  
> I am very interested in the underlying registration technique in ITK. 
> I realize that ITK registration is done in physical coordinates, all 
> the voxels in two datasets should be transformed to points in physical 
> space before registration.
>  
> 1. Generally the number of points in two datasets to be registered are 
> different, are they resampled to the same number of points in ITK 
> registration framework - is there any interpolated point 
> involved during registration? If the two point datasets are in 
> non-uniform grid, how to generate the joint histogram, for example the 
> element h(i, j) in joint histogram represents the number of point 
> pairs which have grey value of i in one dataset, and have grey value 
> of j in the corresponding positions in the other dataset. My question 
> is how to build the corresponding position if the points are on 
> non-uniform grids.

Sure the metrics take that into account..
Take a look at the methods TransformIndexToPhysicalPoint and 
TransformPhysicalPointToContinuousIndex in ImageBase.

they do something like

Iterate over every pixel in the image with an indexed iterator
{
   fixedImage->TransformIndexToPhysicalPoint( fixedImageIndex, inputPoint );
   transformedPoint  = transform->TransformPoint( inputPoint );
   movingImage->TransformPhysicalPointToContinuousIndex( 
transformedPoint, MovingImageIndex );
  
   // Now you have the indices of the correspoinding pixel in the moving 
image..
...

}

>  
> 2. Since ITK take into account the voxel size, and use it to compute a 
> scaling factor for transform, is this scaling factor reflected in the 
> final transform matrix?
>  

Scaling ?? The transform is in physical coordinates. (real world 
coordinates).

> Many thanks in advance!
>  
> Regards,
>  
> Yan
>  
>  
>  
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Insight-users mailing list
>Insight-users at itk.org
>http://www.itk.org/mailman/listinfo/insight-users
>  
>