[Insight-users] Actually implementing a Gabor filter

[Nicholas Tustison]

Can you try the nifti format (*.nii) and see what that does?

On Jun 22, 2010, at 12:34 PM, Neal R. Harvey wrote:

> Nicholas Tustison wrote:
>> Okay, then, let me ask this---have you actually ran the examples untouched (for both contributions) and ensured that the outputs of the examples can be run on your machine and are correct?
>>  
> I have built the examples, successfully. I have managed to create a gabor kernel image using the test code from the "go-go Gabor gadgetry" paper.
> However, I have not had success in using this gabor kernel image with the test code from the "Image kernel convolution" paper. I was very limited
> in the output image type that the test code would let me write the Gabor kernel image. One that the code did allow me to write was a VTKimage.
> 
> command: itkGaborImageSourceTest /tmp/testGaborKernel.vtk 0
> 
> No error messages and the file /tmp/testGaborKernel.vtk was successfully written.
> 
> command: ls -l /tmp/testGaborKernel.vtk
> -rw-rw---- 1 harve harve 262476 Jun 22 10:28 /tmp/testGaborKernel.vtk
> 
> Then, when I use this with the convolution test code, I get the following message:
> 
> command: itkConvolutionImageFilterTest 2 test0.img /tmp/testGaborKernel.vtk /tmp/testGaborOutput.tif
> 
> Message:
> 
> WARNING: In /data0/harve/itk/InsightToolkit-3.18.0/Code/Common/itkImageBase.txx, line 289
> Image (0x910ad48): Not executing UpdateOutputData due to zero pixel condition RequestedRegion:ImageRegion (0x910ae2c)
> Dimension: 2
> Index: [0, 0]
> Size: [0, 0]
> BufferedRegion: ImageRegion (0x910ae40)
> Dimension: 2
> Index: [0, 0]
> Size: [256, 256]