[Insight-users] how to use ReinitializeLevelSetImageFilter
Luis Ibanez
luis.ibanez at kitware.com
Thu Mar 18 12:47:14 EDT 2010
Hi Baoyun,
You may not need to reinitialize the level set at every iteration.
In principle, the zero set doesn't move more than one pixel at
every iteration, so I would think that you should only need to
reinitialize after a number of iterations that is equal to the half
width of the narrow band (measured in pixels).
Whether the approach is practical in term of computation time is
something that can only be defined by knowing the requirements
of your segmentation and the capabilities of the computation
platform that you have at hand.
I would suggest that you instrument the code with TimeProbes
and profile how much time is spent on every subtask.
Please, make sure that you are compiling for Release before
you start profiling the computation time.
Regards,
Luis
----------------------------------------------------------------------------------------------
On Mon, Mar 8, 2010 at 5:49 PM, Baoyun Li <baoyun_li123 at yahoo.com> wrote:
> Hi, Luis:
>
> Thanks so much.
>
> I program is based release model. Five mimutes is for not using narrarow
> band. How long time usually need for this filter with narrow band 12 for the
> 512*512*100 ct images? My plan is to use ReinitializeLevelSetImageFilter at
> each iteration. Do you think it is practial in terms of computation time.
>
> I now make the ReinitializeLevelSetImageFilter works using binary image.
> The image has 0 and 255 intensity value. I set the levesetvale to be 128, it
> seems works.
>
> My best regars
>
> Baoyun
>
> --- On *Sat, 3/6/10, Luis Ibanez <luis.ibanez at kitware.com>* wrote:
>
>
> From: Luis Ibanez <luis.ibanez at kitware.com>
> Subject: Re: [Insight-users] how to use ReinitializeLevelSetImageFilter
> To: "Baoyun Li" <baoyun_li123 at yahoo.com>
> Cc: insight-users at itk.org
> Date: Saturday, March 6, 2010, 10:28 PM
>
>
>
> Hi Baoyun,
>
> The input to the ReinitializeLevelSetImageFilter
> is not expected to be a binary image.
>
> The input image is supposed to be a level set image.
>
> The purpose of the filter is to correct the drift that will
> happens in a level set image due to the inaccuracies
> of computing its progress based only on a narrow band
> of pixels.
>
> Five minutes is an abnormally long time for this filter.
>
> Are you compiling your code for "Debug" mode ?
>
>
> Luis
>
>
>
> ------------------------------------------------------------------------------------------
> On Tue, Mar 2, 2010 at 11:53 AM, Baoyun Li <baoyun_li123 at yahoo.com<http://us.mc1110.mail.yahoo.com/mc/compose?to=baoyun_li123@yahoo.com>
> > wrote:
>
>> Dear Luis:
>>
>> I tried to add SetLevelSetValue(0) and SetLevelSetValue(128). And set
>> narrowbandoff.
>> However, I am getting same weird result.
>>
>> So can you tell me whether the input of ReinitializeLevelSetImageFilter is
>> binary image.
>>
>> And the filter needs five mintues to run, is it normal for a 512x512x60
>> image.
>>
>> Thanks and my best regards
>>
>> Baoyun
>>
>>
>> --- On *Mon, 3/1/10, Luis Ibanez <luis.ibanez at kitware.com<http://us.mc1110.mail.yahoo.com/mc/compose?to=luis.ibanez@kitware.com>
>> >* wrote:
>>
>>
>> From: Luis Ibanez <luis.ibanez at kitware.com<http://us.mc1110.mail.yahoo.com/mc/compose?to=luis.ibanez@kitware.com>
>> >
>> Subject: Re: [Insight-users] how to use ReinitializeLevelSetImageFilter
>> To: "Baoyun Li" <baoyun_li123 at yahoo.com<http://us.mc1110.mail.yahoo.com/mc/compose?to=baoyun_li123@yahoo.com>>
>>
>>
>> Cc: insight-users at itk.org<http://us.mc1110.mail.yahoo.com/mc/compose?to=insight-users@itk.org>
>> Date: Monday, March 1, 2010, 11:50 PM
>>
>>
>>
>> Hi Baoyun,
>>
>>
>> It looks like you missed to call the method:
>>
>> SetLevelSetValue()
>>
>> before you called Update() in the filter.
>>
>> For guidance, please look at the file:
>>
>> Insight/Testing/Code/Algorithms/
>> itkReinitializeLevelSetImageFilterTest.cxx
>>
>>
>> Regards,
>>
>>
>> Luis
>>
>>
>> ------------------------------------------------------
>> On Mon, Mar 1, 2010 at 2:56 PM, Baoyun Li <baoyun_li123 at yahoo.com<http://us.mc1110.mail.yahoo.com/mc/compose?to=baoyun_li123@yahoo.com>
>> > wrote:
>>
>>> Dear All:
>>>
>>> I tried to use ReinitializeLevelSetImageFilter. The output looks weird to
>>> me.
>>>
>>> The input to the filter is binary image binary image with float type
>>> (255 as forground, 0 as background).
>>>
>>> I tried narroow band on and off. However, the output of this filter gives
>>> very high value (10^38) for voxles inside the binary mask. Almost all the
>>> outside mask voxels has value zero.
>>>
>>> Below is my code, can somebody tell where I made mistake.
>>>
>>> Baoyun
>>>
>>>
>>> **************************code************************
>>> typedef typename itk::ReinitializeLevelSetImageFilter<
>>> FeatureImageType > DistanceFilterType;
>>> typename DistanceFilterType::Pointer DistanceFilter =
>>> DistanceFilterType::New();
>>> DistanceFilter->SetInput( thresholder->GetOutput() );
>>> // DistanceFilter->NarrowBandingOn();
>>> // DistanceFilter->SetNarrowBandwidth( 10 );
>>> try
>>> {
>>> DistanceFilter->Update();
>>> }
>>> catch( itk::ExceptionObject & excep )
>>> {
>>> std::cerr << "Exception caught !" << std::endl;
>>> std::cerr << excep << std::endl;
>>> }
>>>
>>> m_levelset2=DistanceFilter->GetOutput();
>>>
>>> ****************code end*************************
>>>
>>>
>>> _____________________________________
>>> 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
>>>
>>>
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20100318/81a2e179/attachment-0001.htm>
More information about the Insight-users
mailing list