[Insight-users] itk::AffineTransform

[Luis Ibanez]

Hi Motes,

You are right:

   An Affine Transform can represent a composition of

             * Translation
             * Rotation
             * Scaling
             * Shearing


So,Yes,
You could use an Affine transform in order to register
two image whose misalignment is composed of a rotation
and a translation.

However, it will be better to use a RigidTransform, since
you don't need the scaling and shearing parts of the
Affine transform.

If you use a RigidTransform in 3D you only need to optimize
a cost function in a 6-dimensional space, while if you use
an Affine transform, you need to optimize a cost function
in a 12-dimensional space.


Note that, sometimes it is advantageos to solve first the
Translational misalignment, and then correct for the rotation.
It all depends on how large either one is.


     Regards,


         Luis


---------------------
motes motes wrote:
> As I understand the itk::AffineTransform is composed of translation, 
> rotation, scaling and shear. So if I have a moving image that is 
> translated a few mm to the left and then rotated a few degrees clockwise 
> I could use itk::AffineTransform once to resolve the misalignment. 
> Alternatively I had to first do a translation transform followed by a 
> rotation transform which yields 2 transforms.
> 
> Is it correct that itk::AffineTransform is basically just used to 
> combine a sequence of transforms (translation, rotation, scaling and 
> shear) into one?
> 
> 
> ------------------------------------------------------------------------
> 
> _____________________________________
> Powered by www.kitware.com
> 
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
> 
> Please keep messages on-topic and check the ITK FAQ at: http://www.itk.org/Wiki/ITK_FAQ
> 
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users