[Insight-developers] GetInverse is unusable -- patch attached

[Pavel Koshevoy]

Hi,

I've talked to my supervisors at the UofU and they agreed that I can contribute
some of our code to ITK.

I've attached all the transforms I've mentioned earlier, except the triangle
mesh transform -- that one will need some cleaning up and untangling from my
helper classes.  I also have an ImageMosaicVarianceMetric for use when
registering several images simultaneously -- there is no fixed image, all images
are warped.  I'd be happy to contribute that as well, but right now it has
external code dependencies which I don't want to drag along.

What is the right way to get this code contributed to ITK?  Can someone help me?

	Pavel.


Rupert Brooks wrote:
> Hi Pavel,
> 
> I really like the idea of having a GetInverse method at the Transform
> Base class level.  It would also be nice to have a Compose() method
> for similar reasons.   Its also clever to make the
> InverseTransformType different from the current transform type
> (obvious in retrospect... but i hadnt thought of it).
> 
> Its of course not up to me whether this gets implemented in the main
> trunk.  Just a quick tip though - I have similar problems with my
> code, but rather than patching ITK which as you point out is quite
> painful for your users, i used a Facade class.  I tucked generic
> Compose and Inverse methods (and misc other things) in there.  Much
> less elegant though, it ends up amounting to a bunch of switch
> statements.  Still, it might get you around your code distribution
> issue in the short term.
> 
> http://en.wikipedia.org/wiki/Facade_pattern
> 
> In my case, my facade class takes a transform in its constructor, and
> then supplies GetInverse(), and Compose methods(), and so forth for
> that transform.
> 
> That being said, the transforms you have created sound extremely
> useful - I for one would be very interested in seeing an Insight
> journal submission or other release.  I had been considering writing
> something similar to your NewtonMethodInverse but never found the
> time.
> 
> Cheers,
> Rupert B.
> 
> 
> On Fri, Oct 10, 2008 at 10:55 AM, Pavel Koshevoy <koshevoy at sci.utah.edu> wrote:
>> Hi,
>>
>> I've been patching ITK since version 1.8.1 to provide a usable GetInverse
>> API for transforms which don't have an analytic inverse, or have an
>> approximate inverse transform of a different type.
>>
>> Requiring a patched ITK is a hurdle in the path of developers trying to
>> build my image registration apps.  I really, really hope you merge this
>> patch into the trunk.
>> It's a very small change.  Basically, I define an InverseTransformPointer
>> type and add virtual InverseTransformPointer GetInverse() const method.  The
>> patch implements this API for itk::Transform, itk::IdentityTransform,
>> itk::MatrixOffsetTransformBase, itk::ScaleTransform,
>> itk::TranslationTransform.
>> These are useful changes that make the ITK Transform classes far more
>> useful. In my image registration apps I implement this API in my Legender
>> Polynomial Transform, a Radial Basis Function Transform, Triangle Mesh
>> Transform, Radial Distortion Transform, Cascaded Transform, and a generic
>> Newtons Method Numeric Inverse Transform.
>>
>> I really hope someone takes a look at this short patch and adds it to the
>> main ITK trunk.
>>
>> Thank you,
>>   Pavel.