[Insight-users] reproducible bug in CannyEdgeDetectionImageFilter

[Luis Ibanez]

Hi Charl,


           That looks like a bug.


The filter should produce identical images,
when called with identical parameters.

Could your please file a bug report in the bug tracker ?

Also, in the meantime, you may want to try the modified
CannyEdge detection filter that we improved for the
Lesion Sizing Toolkit:

https://www.kitware.com/svn/LesionSizingKit/trunk/Sandbox/Source/

You may want to take the files:

itkCannyEdgeDetectionRecursiveGaussianImageFilter2.h<https://www.kitware.com/svn/LesionSizingKit/trunk/Sandbox/Source/itkCannyEdgeDetectionRecursiveGaussianImageFilter2.h>
itkCannyEdgeDetectionRecursiveGaussianImageFilter2.txx<https://www.kitware.com/svn/LesionSizingKit/trunk/Sandbox/Source/itkCannyEdgeDetectionRecursiveGaussianImageFilter2.txx>

Regards,


      Luis


-------------------------------------
On Mon, Aug 10, 2009 at 12:33 PM, Charl Botha <c.p.botha at tudelft.nl> wrote:

> Dear all,
>
> Running the attached Python script on
> Insight/Testing/Data/Input/cthead1.png yields 4 images on disc, of
> which canny_t0.png and canny_t0_after.png should be identical, as they
> are using the same thresholds, and otherwise identical parameters.
> However, they are wildly different, see the attached images.
>
> In short, the canny filter's behaviour changes depending on whether it
> has been freshly instantiated or whether it has been used with a
> different parameter setting before.  Setting the threshold back to a
> previous value after having executed the filter with different values
> will yield incorrect results.
>
> Is this a bug, or am I expecting too much from this filter? :)
>
> Thanks,
> Charl
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.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
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20090815/76ec1be8/attachment.htm>