[Insight-users] bug in MattesMutualInformationImageToImageMetric

[Marius Staring]

Hi Luis,

I did so. The bug number is #3998.

Marius

Luis Ibanez wrote:
>
> Hi Marius,
>
> Thanks for pointing this out.
>
> Could you please enter a bug report in the BugTracker
>
>     http://public.kitware.com/Bug/index.php
>
>
> Please let us know if you find any problem when creating
> the bug report.
>
>
>    Thanks
>
>
>      Luis
>
>
> ----------------------
> Marius Staring wrote:
>> Hi all,
>>
>> I want to report a bug in the 
>> MattesMutualInformationImageToImageMetric. The bug is when using all 
>> the pixels in the fixed image and when a mask is used.
>>
>> In Initialize() memory is allocated for the fixed sample container, 
>> of the size of the fixed image region. Later, in the 
>> SampleFullFixedImageDomain() function the container is filled, but of 
>> coarse, since the number of samples within the mask is usually less 
>> than that of the full fixed image, the fixed sample container is not 
>> completely filled. This results in an access violation when calling 
>> GetValue[AndDerivative]().
>>
>> - In the SampleFullFixedImageDomain() a counter should be used, to 
>> count the actual number of pixels.
>> - The container should be resized accordingly
>> - the m_NumberOfSpatialSamples should be updated
>> - the while loop should also stop at the end of the regioniter, so
>>
>>        while( iter != end )
>>
>> should be something like
>>             while( iter != end && !regioniter.IsAtEnd() )
>>
>> - Also the comments should be updated, since no random iterator is used
>>
>> Hope this helps, cheers
>>
>> Marius
>>
>

-- 
Marius Staring
Image Sciences Institute
University Medical Centre Utrecht
Heidelberglaan 100, 3584 CX Utrecht, The Netherlands
phone: +31 (0)30 250 3186, fax: +31 (0)30 251 3399
marius at isi.uu.nl, http://www.isi.uu.nl/People/Marius