[Insight-developers] RE: CenteredTransformationInitializer

Lorensen, William E (Research) lorensen@crd.ge.com
Fri, 31 Jan 2003 14:57:43 -0500


No problem. I use ChangeInformation now to center by models. I would not bother to try and slip it
in.

Thanks,

Bill

-----Original Message-----
From: Luis Ibanez [mailto:luis.ibanez@kitware.com]
Sent: Friday, January 31, 2003 2:38 PM
To: Lorensen, William E (Research)
Cc: Insight-developers (E-mail)
Subject: Re: CenteredTransformationInitializer


Bill,

That's right,

The Initializer is ND but requires the transform to have
some methods that right now are only available in the
Centered2D transforms.

I plan to add 'Centered' transforms for 3D. The consequence
is that they will increase their parameters array by
3 elements.

Probably the stability they provide for the optimizer
compensates largely this disadvantage.

I may be able to put a couple of centered 3D transforms during
the weekend so they get included in the release.



   Luis


---------------------

Lorensen, William E (Research) wrote:

> Luis,
> The subject filter (residing in Common) uses itkImageMomentsCalculator which resides in Algorithms.
> Should it bemoved to Common?
> 
> Also, I wanted to use the CenteredTransformInitializer to initialize a Rigid3D transform. Of
course,
> it applies the SetCenter and SetTranslation methods to its transform and Rigid3D does not have
those
> methods. I guess it's only valid for 2D transforms?
> 
> 
> Bill
> 
> 
>