[Insight-users] itkImageConvolutionImageFilter/itkGaussianImageSource Problem

[Nicholas Tustison]

Hi Luis,

I'm not quite sure what the correct behavior should be in this case.  Do we 
resample to get an odd radius?

Nick





On May 29, 2011, at 5:41 PM, Luis Ibanez wrote:

> Hi Krishnan,
> 
> You found indeed a bug in the itkConvolutionImageFilter class.
> 
> We just logged a bug entry in the tracker:
> http://public.kitware.com/Bug/view.php?id=12236
> 
> The attached file, (a modified version of yours)
> more clearly illustrates the problem.
> 
> When using a Kernel of even size (e.g. 32) the
> output convolved image has a diagonal shift.
> 
> 
> This should be fixed soon in ITK4.
> 
> 
>    Thanks
> 
> 
>         Luis
> 
> 
> ----------------------------------------
> On Thu, May 26, 2011 at 9:53 AM, Krishnan V <v.srikrishnan at gmail.com> wrote:
>> 
>> 
>> On Thu, May 26, 2011 at 7:08 PM, Neil Panjwani <paniwani at gmail.com> wrote:
>>> 
>>> What are the dimensions of your kernel? They should be small like 3x3 in
>>> the example...
>> 
>> The kernel is much larger. i used the default of the itkGaussianImageSource
>> which gives a 32x32 or 64x64, certainly not 3x3.
>> Also, the even size of the kernel was not expected...
>>> 
>>> 
>>> On Thu, May 26, 2011 at 9:33 AM, Krishnan V <v.srikrishnan at gmail.com>
>>> wrote:
>>>> 
>>>> 
>>>> On Thu, May 26, 2011 at 6:58 PM, Neil Panjwani <paniwani at gmail.com>
>>>> wrote:
>>>>> 
>>>>> Not sure if you've seen this already, but it's a simpler example that
>>>>> should work:
>>>>> 
>>>>> http://www.itk.org/Wiki/ITK/Examples/ImageProcessing/ConvolutionImageFilter
>>>> 
>>>> Thanks for the reply. that works... my example does not.
>>>>> 
>>>>> On Thu, May 26, 2011 at 5:04 AM, Krishnan V <v.srikrishnan at gmail.com>
>>>>> wrote:
>>>>>> 
>>>>>> Hi all,
>>>>>> I am trying out the convolution image filter by generating a simple
>>>>>> Gaussian and convolving it with itself, ie, convolve a gaussian with itself.
>>>>>> This is done in 2D. I get an output which consists of two peaks at two ends
>>>>>> of the image along the X-axis.  This is obviously not the expected output. I
>>>>>> have attached a simple sample code.
>>>>>> Can someone faced a similar issue or help out with this?
>>>>>> thanks
>>>>>> krishnan
>>>>>> 
>>>>>> /************************************************************************************************************
>>>>>> */
>>>>>> #include <itkImage.h>
>>>>>> #include <itkGaussianImageSource.h>
>>>>>> #include <itkConvolutionImageFilter.h>
>>>>>> 
>>>>>> typedef float PixelType;
>>>>>> typedef itk::Image<PixelType, 2> ImageType2D;
>>>>>> typedef itk::ImageFileReader<ImageType2D> Reader2D;
>>>>>> typedef itk::ImageFileWriter<ImageType2D> Writer2D;
>>>>>> void convolve_test(ImageType2D::Pointer input_image,
>>>>>> ImageType2D::Pointer kernel)
>>>>>> {
>>>>>>     typedef itk::ConvolutionImageFilter<ImageType2D,ImageType2D>
>>>>>> Convolver_Type;
>>>>>>     Convolver_Type::Pointer convolve = Convolver_Type::New();
>>>>>> 
>>>>>>     convolve->SetInput(kernel);
>>>>>>     convolve->SetImageKernelInput(kernel);
>>>>>>     convolve->Update();
>>>>>> 
>>>>>>     Writer2D::Pointer imwriter = Writer2D::New();
>>>>>>     imwriter->SetInput(convolve->GetOutput());
>>>>>>     imwriter->SetFileName("lenna.nii");
>>>>>>     imwriter->Update();
>>>>>>     return;
>>>>>> }
>>>>>> int main(int argc, char **argv)
>>>>>> {
>>>>>>   Writer2D::Pointer writer = Writer2D::New();
>>>>>>   ImageType2D::Pointer kernel;
>>>>>>   typedef itk::GaussianImageSource<ImageType2D>
>>>>>> Gaussian_Generator_Type;
>>>>>>   Gaussian_Generator_Type::Pointer kernel_generator =
>>>>>> Gaussian_Generator_Type::New();
>>>>>>   double sigma[3];
>>>>>>   unsigned long size[3];
>>>>>>   double mean[3];
>>>>>> 
>>>>>>   Reader2D::Pointer imreader = Reader2D::New();
>>>>>>   imreader->SetFileName(argv[1]);
>>>>>>   imreader->Update();
>>>>>> 
>>>>>>   sigma[0] = 10.00; sigma[1] = 10.0;
>>>>>>   kernel_generator->SetSigma(sigma);
>>>>>> 
>>>>>>   kernel_generator->Update();
>>>>>>   kernel = kernel_generator->GetOutput();
>>>>>>   convolve_test( imreader->GetOutput(), kernel);
>>>>>>   /* Write the Gaussian */
>>>>>>   writer->SetFileName("Gaussian-1.nii");
>>>>>>   writer->SetInput(kernel);
>>>>>>   writer->Update();
>>>>>> 
>>>>>>   return 0;
>>>>>> }
>>>>>> 
>>>>>> _____________________________________
>>>>>> 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://www.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-users
>>>>>> 
>>>>> 
>>>> 
>>> 
>> 
>> 
>> _____________________________________
>> 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://www.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-users
>> 
>> 
> <convolve.cxx>