[Insight-developers] RecursiveMultiResolutionPyramidImageFilter -- questions about fixing bug 8482 : If it is not tested, then it is....

[Luis Ibanez]

    Yes,
    We need to raise code coverage,
    and add more tests.

    Anything below 100% is unacceptable.

    Even if we had a 100% code coverage,
    we still have to exercise multiple
    combinations of filter settings.


        Luis


----------------------------
Bradley Lowekamp wrote:
> 
> It is unfortunate, but a fact of ITK that the comments and the tests 
> don't cover all of the expected behavior. I think the fact that this bug 
> got through is an clear indication that we need another registration 
> test which is noisy and has non-one spacing.
> 
> Kent, thank you for uncovering this bug. Making the change to not use 
> image spacing to my local copy restored the quality of my registration.
> 
> Brad
> 
>