[Insight-users] direction of Gradient

[Neilson Mackay]

It was based on a problem I am having.

I have mad ea modified metric that takes a set of points and multiple  
3d images.  Each image is associated with a transform that transform  
the image somewhere in 3d space.

Basically I have 1 Registration transform and N image transforms ( n  
= nmber of images):

Metric::GetValue(transform registration_transform)
{
	for each point p in Pointset
	{
		for each image Im in set of images ImageSet
		{
			registration_transform.TransformPoint(p);
			im->getTransform->TransformPoint(p);

			... normal metric stuff
		}
	}
}
	
The problem I am hainvg is I am not sure how to calculate the  
derivitive.  I assume in normal registration, the "jacobian"  
transforms the gradient.  If I have two transforms should I  
concatenate them together then create a jacobain from that?  Or  
multiply the seprate jacobains together?

Does this make sence?  :)  Thanks for the help!

Neilson


On Jul 10, 2008, at 5:41 AM, Luis Ibanez wrote:

>
> Hi Neilson,
>
>                It "should" be independent.
>
> But this is something for which we don't have a
> Nightly test as verification...
>
>
>    Is your question a theoretical one ?
>
>    or do you have a specific ITK filter in mind ?
>
>
>   Please let us know,
>
>
>      Thanks
>
>
>         Luis
>
>
> ----------------------
> Neilson Mackay wrote:
>> Hi,
>> Is the rotation component of a Gradient dependent on the origin  
>> of  the image?  Or is completely independent?
>> Neilson
>> _______________________________________________
>> Insight-users mailing list
>> Insight-users at itk.org
>> http://www.itk.org/mailman/listinfo/insight-users