[Insight-users] Results of itkCurvesLevelSetImageFilter

[Luis Ibanez]

Hi Pingkun,

Thanks for posting the images.

The edgePotential.png clearly explains the behavior
of your LevelSet.

You are passing the negative of what the speed image
should be.  The speed image is supposed to be dark
in the edges and bright inside the object.

With your current speed image what you are doing is
making the little circle grow "inside" the edges and
make it collapse insde the tubular structure that you
were segmenting.

Please invert the intensities of the speed image.

You should read the Chapter on LevelSet segmentation
in the ITK Software Guide


       http://www.itk.org/ItkSoftwareGuide.pdf


Let us know if you have further questions.


    Thanks


       Luis



--------------------
Pingkun Yan wrote:

> Hi Luis,
> 
> Please see my response below.
> 
> Thanks!
> Pingkun
> 
>>
>> Hi Pingkun,
>>
>> Thanks for the clarification.
>>
>> Please post to the list the following:
>>
>> 1) The speed (feature) image that you passed to the Curves filter.
> 
> 
> The speed image "edgePotential.png" is attached.
> 
>>
>> 2) The "circle" image that you generated with FastMarching.
> 
> 
> The image "initialLevelSet.png" is generated by fast marching, where I 
> set the seed value as -3.5.
> 
>>
>> 3) The parameter setting that you select when running the Curves
>>    filter (eg. max number of iterations, max RMS error. etc).
> 
> 
> The parameter setting is:
>  curvesFilter->SetPropagationScaling( 1.0 );
>  curvesFilter->SetCurvatureScaling( 0.1 );
>  curvesFilter->SetAdvectionScaling( 0.5 );
>  curvesFilter->SetMaximumRMSError( 0.03 );
>  curvesFilter->SetNumberOfIterations( 500 );
> 
>>
>> 4) The number of iterations that the filter actually used by
>>    the time it stopped.
> 
> 
> The program prints out following information when finish:
>    No. elpased iterations: 446
>    RMS change: 0.0298426
> 
>>
>>
>> Thanks
>>
>>
>>     Luis
>>
>>
>>
>> ------------------
>> Pingkun Yan wrote:
>>
>>> Hi Luis,
>>>
>>> Thank you for your reply. Actually, I applied the Sigmoid of the 
>>> Gradient Magnitude on the "test.png" to get feature image. Then I 
>>> used FastMarching to generate an initial circle as the test example.
>>>
>>> Thanks!
>>>
>>> Pingkun.
>>>
>>> ----- Original Message ----- From: "Luis Ibanez" 
>>> <luis.ibanez at kitware.com>
>>> To: "Pingkun" <pingkun at ieee.org>
>>> Cc: <insight-users at itk.org>
>>> Sent: Wednesday, September 29, 2004 11:22 PM
>>> Subject: Re: [Insight-users] Results of itkCurvesLevelSetImageFilter
>>>
>>>
>>>>
>>>> Hi Pingkun,
>>>>
>>>> When you passed the image "test.png" as input to the LevelSet filter,
>>>> did you pass it as the "InputImage" or as the "FeatureImage" ?
>>>>
>>>> The example that you mention is building its own Feature and Input
>>>> image. In order to use this level set filter you must provide *both*
>>>> an "Input" image, which is a level set initialization, and a "Feature"
>>>> image, which is typically a speed image, like the one you can create
>>>> with the Sigmoid of the Gradient Magnitude.
>>>>
>>>> Please let us know more details on how did you feed the inputs to
>>>> the filter.
>>>>
>>>>
>>>>   Thanks
>>>>
>>>>
>>>>     Luis
>>>>
>>>>
>>>> ----------------
>>>> Pingkun wrote:
>>>>
>>>>> Dear all,
>>>>>
>>>>> I tried the itkCurvesLevelSetImageFilter on the attached image
>>>>> "test.png" and the results is shown in "outputLevelSet.png". The 
>>>>> results
>>>>> look like two vessels with the true boundaries as their centerlines. I
>>>>> just followed the test file "itkCurvesLevelSetImageFilterTest.cxx".
>>>>> Anybody can point out the problem or give me an example? Thanks!
>>>>>
>>>>> Rgds,
>>>>> Pingkun
>>>>> ------------------------------------------------------------------------ 
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------ 
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------ 
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Insight-users mailing list
>>>>> Insight-users at itk.org
>>>>> http://www.itk.org/mailman/listinfo/insight-users
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Insight-users mailing list
>>> Insight-users at itk.org
>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>>
>>
>>
>>
>>
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users