[Insight-developers] Proposed KernelTransform API Change

[Stephen R. Aylward]

Using FixedParameters seems like the right thing for this!   Works for us...

Thanks,
Stephen


Vincent A. Magnotta wrote:
> Stephen,
> 
> My original proposal was to change the API for SetParameters() to return
> both the source and target landmarks. However, maybe a better solution
> would be to make the  Get and Set methods for the FixedParameters to
> return the the target landmarks. Since these functions do not currently
> exist for the Kernel Transform, no API changes would be required. This
> would also allow the Kernel Transform to support the TransformFileReader
> and TransformFileWriter I/O mechanism. 
> 
> Vince
> 
> 
> 
> On Mon, 2005-10-24 at 11:14, Stephen R. Aylward wrote:
> 
>>Hi,
>>
>>We need to talk about this a bit.   We use these classes heavily in 
>>several of our applications.   We have already developed IO methods for 
>>them.
>>
>>Can you provide more details on what you would like to do?   Where will 
>>the API changes be?   What will be the changes to parameter ordering etc?
>>
>>Thanks,
>>Stephen
>>
>>Vincent A. Magnotta wrote:
>>
>>>I am looking into making the KernelTransform classes work with the
>>>Transform I/O mechanism. Currently, these functions only set the source
>>>landmarks as the parameters. This does not completely specify the
>>>transform. I would propose that the parameters specify both the source
>>>and target landmarks. The current approach seems to done to support
>>>optimizers in the registration process. I have looked and could not find
>>>any use of this functionality. If anyone has any thoughts on this, feel
>>>free to let me know. I will make this change by Wednesday unless there
>>>are objections.
>>>
>>>Vince
>>>
>>>

-- 
===========================================================
Dr. Stephen R. Aylward
Associate Professor of Radiology
Adjunct Associate Professor of Computer Science and Surgery
http://caddlab.rad.unc.edu
aylward at unc.edu
(919) 966-9695