[Insight-users] FW: FloodFilledFunctionConditionalConstIterator neighborhood

[Luis Ibanez]

Hi Hauke,

It is great to hear that you have already made the changes.

The procedure for contribute new classes to ITK requires
that they be submitted to the Insight Journal:

http://www.itk.org/Wiki/ITK_Procedure_for_Contributing_New_Classes_and_Algorithms


Note that papers to the Insight Journal *are not typical papers*.
Instead, they follow the format of a technical report. Your paper
may be as short as two pages, as long as it describes how to use
the new class.

What is fundamental is that you post it with an example (or test)
on how to use the class, so that it use can be verified by reviewers
and by readers.


The purpose of the Insight Journal is to *verify reproducibility*.


Once your paper has been reviewed by the community, the code will
be moved into the Insight/Code/Review directory, where it will
be verified for coding style and correctness.


Note for example, that it is a matter of concern, the fact that
you have not verified the correctness of the output, and that
there is not a profiling indicating how the computation time
of the iterator has been affected by the change. Adding features
and functionalities is always a nice thing to do, but in the context
of a toolkit, we should be careful with the implications of the
changes.


All of this is necessary due to the backward compatibility policy
of ITK, that states that once a class has been included with a
release, its API must be maintained *forever*.  As a consequence,
we must review very well all code contributions before they made
their way into the toolkit.


It will be great if you could submit your code as a new class
to the Insight Journal.


   Please let us know if you have any questions,


      Thanks


         Luis


-------------------
Hauke Heibel wrote:
> Hi Luis,
> 
> I have now a version that uses a ShapedNeighborhoodIterator. Per default the
> FloodFilledFunctionConditionalConstIterator performs iteration in the same
> way as the original iterator though it offers now methods to Set/Get a
> parameter called m_FullyConnected. I used the implementation of
> setConnectivity(...) from itkConnectedComponentAlgorithm.h to setup the
> ShapedNeighborhoodIterator.
> 
> The new code is attached to this mail and should integrate and work out of
> the box though I have to mention that I did not check, if by chance the
> order in which points are processed changed - this might be the case if the
> ShapedNeighborhoodIterator computes the offset value in a different way than
> the old filter.
> 
> By the way, is the only way to make even such minor changes to existing code
> become integrated into the repository by filing a complete Journal
> Publication? If that really is the case I will follow the rules and do it
> but I was just wondering if e.g. a feature request on mantis would already
> be sufficient (I could attach the code as well as a CVS patch).
> 
> Best regards,
> Hauke
> 
> -----Original Message-----
> From: Luis Ibanez [mailto:luis.ibanez at kitware.com] 
> Sent: Sunday, November 25, 2007 3:27 PM
> To: Hauke Heibel
> Cc: insight-users at itk.org
> Subject: Re: [Insight-users] FW: FloodFilledFunctionConditionalConstIterator
> neighborhood
> 
> 
> Hi Hauke,
> 
> Yes the FloodFilledFunctionConditionalConstIterator uses 4-connectivity
> when used in 2D images, and, in general, it uses face-connectivity in
> N-D images.
> 
> If you wanted to modify this behavior, the key place in the code is
> in lines 175-178 of
> 
> 
>    Insight/Code/Common/
>       itkFloodFilledFunctionConditionalConstIterator.txx
> 
> 
> It is suggested the code, that the existing for loop could be replaced
> with a ShapeNeighborhoodIteration, in order to provide flexibility
> in the connectivity.
> 
> 
>      That will make an interesting course project,
>      and...
>      a nice paper for the Insight Journal   :-)
> 
> 
> 
>    Regards,
> 
> 
>      Luis
> 
> 
> -----------------
> Hauke Heibel wrote:
> 
>>Hi everybody,
>>
>> 
>>
>>A colleague and me are wondering if it is true that the 
>>FloodFilledFunctionConditionalConstIterator uses a hard coded 
>>4-neighborhood (in 2D)?
>>
>> 
>>
>>If that is the case does anybody already have a version at hand in which 
>>the neighborhood can be defined in a more generic way?
>>
>> 
>>
>>Any help is highly appreciated!
>>
>> 
>>
>>Best regards,
>>
>>Hauke
>>
>> 
>>
>>p.s. I hope the message will not arrive twice since I sent it through 
>>the wrong account in the beginning.
>>
>>
>>------------------------------------------------------------------------
>>
>>_______________________________________________
>>Insight-users mailing list
>>Insight-users at itk.org
>>http://www.itk.org/mailman/listinfo/insight-users
>>
>>
>>------------------------------------------------------------------------
>>
>>_______________________________________________
>>Insight-users mailing list
>>Insight-users at itk.org
>>http://www.itk.org/mailman/listinfo/insight-users