[Insight-developers] Adding Consistency check to itk::BinaryFunctorImageFilter : Design Question

[Luis Ibanez]

Gaetan,

That's a very good point.

The method should indeed be available at a higher level
in the filter hierarchy. Probably at the level of the

                 itkImageToImageFilter

How about offering a method that does:

    bool CheckImageInformationConsistency(
         const ImageBase * image1,
         const ImageBase * image2
         ) const;

This could still use helper method for specific
comparisons, such as

    bool CheckImageSpacingConsistency(
         const ImageBase * image1,
         const ImageBase * image2
         ) const;

...

 bool CheckImageOriginConsistency( ...)
 bool CheckImageDirectionConsistency(...)

....


Looking at what we are comparing, this is all
information that is available at the level of the
ImageBase, therefore we may manage to base
it only on the image dimension.

Filters down the hierarchy could use these
methods in different combinations, according
to what applies to their particular cases.

BTW: one concern is that the Origin, Spacing
and Direction comparisons may have to be
made with a certain numerical tolerance.



What do you think ?


       Thanks


               Luis



-----------------------------------------------------
2009/12/2 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>
> Hi Luis,
>
> I personally prefer your first approach. Don't you think that
> CheckInputImagesConsistency() should be in BeforeThreadedGenerateData() in
> BinaryFunctorImageFilter directly, instead of its subclasses? The tests may
> be desactivated in the subclasses if needed, by reimplementing
> BeforeThreadedGenerateData().
>
> Also, this bug must also be in many other classes which are processing
> several inputs - maybe those methods should be added in a more basic class,
> so they can be reused in those filters as well?
>
> Gaëtan
>
>
> Le 2 déc. 09 à 21:12, Luis Ibanez a écrit :
>
>> The Bug 9893:
>>
>>           http://www.itk.org/Bug/view.php?id=9893
>>
>> calls for adding a check of consistency between the two input
>> images of filters that derive from itkBinaryFunctorImageFilter.
>>
>> The example, is the AddImageFilter, in which we implicitly
>> assume that the two images passed as input have:
>>
>>   1) The same number of pixels along each dimension
>>   2) The same origin
>>   3) The same spacing
>>   4) The same direction
>>
>>
>> Currently there is no check for these assumption,
>> and therefore, a user can pass two images of different
>> sizes, and get a segmentation fault at run time.
>>
>> ---
>>
>> The proposed fix is to add the following methods:
>>
>>>  /** These method verify if the two input images have the same
>>>  *  size, origin, spacing and orientation. It is intended to be
>>>  *  called from some derived filters, when they require the two
>>>  *  inputs to be consistent. The method throws and exception if
>>>  *  the two input images are not consistent. */
>>>  void CheckInputImagesConsistency();
>>>  void CheckInputImagesConsistencyInSize();
>>>  void CheckInputImagesConsistencyInOrigin();
>>>  void CheckInputImagesConsistencyInSpacing();
>>>  void CheckInputImagesConsistencyInDirection();
>>
>>
>> to the protected section of the itkBinaryFunctorImageFilter.
>>
>>
>> Then, most of its derived classes will call the method:
>>
>>           CheckInputImagesConsistency()
>>
>>
>> (which internally will call all the other methods, each one
>> of which could throw an exception if the consistency check
>> fails).
>>
>>
>> Particular filters that do not require all checks, could call
>> only those that apply.
>>
>> The suggested place to make these calls is the method
>>
>>                BeforeThreadedGenerateData();
>>
>>
>> ---
>>
>> An alternative approach is to set these calls in the
>> itkBinaryFunctorImageFilter, and add a boolean that
>> sets whether the check should be performed or not.
>>
>> Derived filters could then set that boolean in their
>> constructors.
>>
>> ----
>>
>>
>> Any preferences between the two approaches above ?
>>
>> Any objections ?
>>
>> or Backward Compatibility concerns ?
>>
>>
>>    Thanks
>>
>>
>>         Luis
>> _______________________________________________
>> Powered by www.kitware.com
>>
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>>
>> Kitware offers ITK Training Courses, for more information visit:
>> http://kitware.com/products/protraining.html
>>
>> Please keep messages on-topic and check the ITK FAQ at:
>> http://www.itk.org/Wiki/ITK_FAQ
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.itk.org/mailman/listinfo/insight-developers
>
> --
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr  http://www.itk.org
> http://www.mandriva.org  http://www.bepo.fr
>
>