[Insight-users] 3D segmentation of LUNGS ( Region growing algorithm ) Need help
Luis Ibanez
luis.ibanez at kitware.com
Thu Mar 12 15:46:45 EDT 2009
Hi Wassim,
If you are not getting the entire structure that you are looking for,
then typical solutions are:
A) Lower the number of your Low Threshold
or
B) Raise the number of your Upper Threshold
or
C) Add more seed points, in particular, in the
locations that your segmentation is currently missing.
Regards,
Luis
--------------------------------
wassim_belhadj at topnet.tn wrote:
> Hi,
>
> Thank you very much for your help.
>
> I was able to do the segmentation on an "unsigned char" raw file ( 512 *
> 512 * 494 ).
>
> The filter gave me a binary mask of the two lungs.
>
> I tried to apply the same algorithm to separate 2D dicom (512 *512) that
> form the Raw file.
>
> These Dicom file are encoded in 16 bit (signed short).
>
> I changed only the lowerThreshold (-950) and the upperThreshold (-550) and
> SetReplaceValue to 1100.
>
> I have not had the same result. Some images contain only the left side of
> the lungs.
>
> That was not the case on the images coded on 8 bit.
>
> I doubt that itk::Image does not support "signed short" type.
>
>
> Can you help me please.
>
> Thank you in advance
>
> /***********************/
> typedef signed short InternalPixelType;
> typedef signed short OutputPixelType;
> const unsigned int Dimension = 2 ;
>
> typedef itk::Image< InternalPixelType, Dimension > InternalImageType;
> typedef itk::Image< OutputPixelType, Dimension > OutputImageType;
>
> ....
> /***********************/
>
>
>
>
> On Mon, 23 Feb 2009 07:29:43 -0500, Luis Ibanez <luis.ibanez at kitware.com>
> wrote:
>
>>Hi Wassim,
>>
>>
>>The error below simply means that the type of ImageFileWriter
>>that you are instantiating doesn't match the output image type
>>of the the connectedThreshold filter.
>>
>>
>>Please change the declaration:
>>
>> >> > typedef itk::ConnectedThresholdImageFilter< InternalImageType,
>> >> > InternalImageType > ConnectedFilterType;
>>
>>
>>to
>>
>>
>> >> > typedef itk::ConnectedThresholdImageFilter< InternalImageType,
>> >> > OutputImageType > ConnectedFilterType;
>>
>>
>>So that the output type of the region growing filter match
>>the expected input type of the writer.
>>
>>
>>
>> Regards,
>>
>>
>> Luis
>>
>>
>>
>>--------------------------------
>>wassim_belhadj at topnet.tn wrote:
>>
>>>Hi Luis,
>>>
>>>You said me that I don't need a Cast filter.
>>>
>>>If I don't use a CastingFilterType::Pointer I have this Error message
>>>at
>>>line :
>>>
>>>writer->SetInput( connectedThreshold->GetOutput() )
>>>
>>>cannot convert parameter 1 from 'itk::Image<TPixel,VImageDimension> *'
>
> to
>
>>>'const itk::Image<TPixel,VImageDimension> *'
>>>
>>>I am using ImageJ and trial Volview version for looking at the output
>>>image.I have got the same result.
>>>
>>>
>>>I located my seed points on a 2D Slice ( lungs parts).
>>>
>>>Thank you in advance
>>>
>>>
>>>
>>>On Fri, 20 Feb 2009 12:30:19 -0500, Luis Ibanez
>
> <luis.ibanez at kitware.com>
>
>>>wrote:
>>>
>>>
>>>>Hi Wassim,
>>>>
>>>>
>>>> Important Rule: Start Simple
>>>>
>>>>
>>>>You don't need the Smoothing filter (until proven otherwise),
>>>>You don't need the Cast filter
>>>>You don't need the ITK-VTK import export code.
>>>>
>>>>
>>>>
>>>>You code could be reduced to:
>>>>
>>>>
>>>>
>>>>>#include "itkConnectedThresholdImageFilter.h"
>>>>>#include "itkImage.h"
>>>>>#include "itkImageFileReader.h"
>>>>>#include "itkImageFileWriter.h"
>>>>>
>>>>>int main( int argc, char *argv[])
>>>>>{
>>>>>
>>>>> typedef signed short InternalPixelType;
>>>>> typedef unsigned char OutputPixelType;
>>>>> const unsigned int Dimension = 3;
>>>>> typedef itk::Image< InternalPixelType, Dimension >
>>>>
>>>> InternalImageType;
>>>>
>>>>> typedef itk::Image< OutputPixelType, Dimension > OutputImageType;
>>>>>
>>>>> typedef itk::ImageFileReader< InternalImageType > ReaderType;
>>>>> typedef itk::ImageFileWriter< OutputImageType > WriterType;
>>>>>
>>>>> ReaderType::Pointer reader = ReaderType::New();
>>>>> WriterType::Pointer writer = WriterType::New();
>>>>>
>>>>> reader->SetFileName( "c:/VTKEdge_Base_Test/images/Input.mhd" );
>>>>> writer->SetFileName( "c:/VTKEdge_Base_Test/images/Output.mhd" );
>>>>>
>>>>>
>>>>> typedef itk::ConnectedThresholdImageFilter< InternalImageType,
>>>>>InternalImageType > ConnectedFilterType;
>>>>>
>>>>> ConnectedFilterType::Pointer connectedThreshold =
>>>>>ConnectedFilterType::New();
>>>>>
>>>>> connectedThreshold->SetInput( reader->GetOutput() );
>>>>> writer->SetInput( connectedThreshold->GetOutput() );
>>>>>
>>>>>
>>>>> const InternalPixelType lowerThreshold = -700 ;
>>>>> const InternalPixelType upperThreshold = 200 ;
>>>>>
>>>>> connectedThreshold->SetLower( lowerThreshold );
>>>>> connectedThreshold->SetUpper( upperThreshold );
>>>>>
>>>>> connectedThreshold->SetReplaceValue( 255 );
>>>>>
>>>>> InternalImageType::IndexType index;
>>>>> index[0] = 180;
>>>>> index[1] = 296 ;
>>>>>
>>>>> connectedThreshold->SetSeed( index );
>>>>>
>>>>> InternalImageType::IndexType index2;
>>>>> index2[0] = 153;
>>>>> index2[1] = 274;
>>>>>
>>>>>
>>>>> connectedThreshold->SetSeed( index2 );
>>>>>
>>>>> try
>>>>> {
>>>>> writer->Update();
>>>>> }
>>>>> catch( itk::ExceptionObject & excep )
>>>>> {
>>>>> std::cerr << "Exception caught !" << std::endl;
>>>>> std::cerr << excep << std::endl;
>>>>> }
>>>>>
>>>>> return 0;
>>>>>}
>>>>
>>>>
>>>>
>>>>
>>>>BTW: in your code below you were calling SetInput()
>>>> in the threshold filter twice, with two different
>>>> inputs. Only the second would take place. This is
>>>> fixed in the code above.
>>>>
>>>>
>>>>Also, please let us know what visualization tool
>>>> you are using for looking at the output image.
>>>> That is: How are you arriving to the conclusion
>>>> that the output image is all set to
>>>> zero values.
>>>>
>>>>
>>>>If the image is actually all set to zero values, then
>>>>the most common reason is that your seed points are
>>>>located in incorrect places (e.g. in places whose
>>>>intensity is not in the range -700HU to 200HU.
>>>>
>>>>
>>>>
>>>> Regards,
>>>>
>>>>
>>>>
>>>> Luis
>>>>
>>>>
>>>>
>>>>------------------------------------
>>>>wassim_belhadj at topnet.tn wrote:
>>>>
>>>>
>>>>>Hi Luis,
>>>>>
>>>>>Thank you very much for your help.
>>>>>
>>>>>I just followed the steps you gave me. It's a bit difficult for a
>>>>>beginner
>>>>>ITK user.
>>>>>
>>>>>
>>>>>The segmentation was not conclusive (A black Volume result).
>>>>>
>>>>>
>>>>>can you help me ?
>>>>>
>>>>>
>>>>>Below my Input.mhd :
>>>>>
>>>>>
>>>
>>>
> /***************************************************************************************************/
>
>>>
> /***************************************************************************************************/
>
>>>>>NDims = 3
>>>>>DimSize = 512 512 494
>>>>>ElementSpacing =0.777344 0.777344 0.7
>>>>>Position = 0 0 0
>>>>>ElementByteOrderMSB = False
>>>>>ElementType = MET_SHORT
>>>>>HeaderSize = -1
>>>>>ElementDataFile = prodhom.raw
>>>>>
>>>>>
>>>
>>>
> /***************************************************************************************************/
>
>>>
> /***************************************************************************************************/
>
>>>>>This is the code that I used :
>>>>>
>>>>>
>>>
>>>
> /***************************************************************************************************/
>
>>>
> /***************************************************************************************************/
>
>>>>>#include "itkConnectedThresholdImageFilter.h"
>>>>>#include "itkImage.h"
>>>>>#include "itkCastImageFilter.h"
>>>>>#include "itkCurvatureFlowImageFilter.h"
>>>>>#include "itkImageFileReader.h"
>>>>>#include "itkImageFileWriter.h"
>>>>>#include "itkVTKImageExport.h"
>>>>>#include "itkVTKImageImport.h"
>>>>>
>>>>>int main( int argc, char *argv[])
>>>>>{
>>>>>
>>>>> typedef signed short InternalPixelType;
>>>>> typedef signed short OutputPixelType;
>>>>> const unsigned int Dimension = 3;
>>>>> typedef itk::Image< InternalPixelType, Dimension >
>
> InternalImageType;
>
>>>>> typedef itk::Image< OutputPixelType, Dimension > OutputImageType;
>>>>> typedef itk::CastImageFilter< InternalImageType, OutputImageType
>>>>>
>>>>>
>>>>>
>>>>>>CastingFilterType;
>>>>>
>>>>> CastingFilterType::Pointer caster = CastingFilterType::New();
>
>
>>>
>>>>>
>>>>> typedef itk::ImageFileReader< InternalImageType > ReaderType;
>>>>> typedef itk::ImageFileWriter< OutputImageType > WriterType;
>>>>>
>>>>> ReaderType::Pointer reader = ReaderType::New();
>>>>> WriterType::Pointer writer = WriterType::New();
>>>>>
>>>>> reader->SetFileName( "c:/VTKEdge_Base_Test/images/Input.mhd" );
>>>>> writer->SetFileName( "c:/VTKEdge_Base_Test/images/Output.mhd" );
>>>>>
>>>>> typedef itk::CurvatureFlowImageFilter< InternalImageType,
>>>>>InternalImageType > CurvatureFlowImageFilterType;
>>>>>
>>>>> CurvatureFlowImageFilterType::Pointer smoothing =
>>>>>CurvatureFlowImageFilterType::New();
>>>>>
>>>>> typedef itk::ConnectedThresholdImageFilter< InternalImageType,
>>>>>InternalImageType > ConnectedFilterType;
>>>>>
>>>>> ConnectedFilterType::Pointer connectedThreshold =
>>>>>ConnectedFilterType::New();
>>>>>
>>>>> smoothing->SetInput( reader->GetOutput() );
>>>>> connectedThreshold->SetInput( smoothing->GetOutput() );
>>>>> connectedThreshold->SetInput( reader->GetOutput() );
>>>>> caster->SetInput( connectedThreshold->GetOutput() );
>>>>> writer->SetInput( caster->GetOutput() );
>>>>>
>>>>> smoothing->SetNumberOfIterations( 5 );
>>>>> smoothing->SetTimeStep( 0.125 );
>>>>>
>>>>> const InternalPixelType lowerThreshold = -700 ;
>>>>> const InternalPixelType upperThreshold = 200 ;
>>>>>
>>>>> connectedThreshold->SetLower( lowerThreshold );
>>>>> connectedThreshold->SetUpper( upperThreshold );
>>>>>
>>>>> connectedThreshold->SetReplaceValue( 255 );
>>>>>
>>>>> InternalImageType::IndexType index;
>>>>> index[0] = 180;
>>>>> index[1] = 296 ;
>>>>>
>>>>> connectedThreshold->SetSeed( index );
>>>>>
>>>>> InternalImageType::IndexType index2;
>>>>> index2[0] = 153;
>>>>> index2[1] = 274;
>>>>>
>>>>>
>>>>> connectedThreshold->SetSeed( index2 );
>>>>>
>>>>> try
>>>>> {
>>>>> writer->Update();
>>>>> }
>>>>> catch( itk::ExceptionObject & excep )
>>>>> {
>>>>> std::cerr << "Exception caught !" << std::endl;
>>>>> std::cerr << excep << std::endl;
>>>>> }
>>>>>
>>>>> return 0;
>>>>>}
>>>>>
>>>
>>>
> /***************************************************************************************************/
>
>>>
> /***************************************************************************************************/
>
>>>
> /***************************************************************************************************/
>
>>>>>Best regards, Wassim
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>On Wed, 18 Feb 2009 10:20:55 -0500, Luis Ibanez
>>>
>>><luis.ibanez at kitware.com>
>>>
>>>>>wrote:
>>>>>
>>>>>
>>>>>
>>>>>>Hi Wassim,
>>>>>>
>>>>>>Please do the following:
>>>>>>
>>>>>>0) Extract a region of interest that includes the Trackea,
>>>>>> and the full Lungs, but not the head of the patient
>>>>>> (e.g. cut below the head, somewhere in the neck).
>>>>>> You can use the RegionOfInterestImageFilter for this.
>>>>>>
>>>>>
>>>>>
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1RegionOfInterestImageFilter.html
>
>>>>>>1) Use the ConnectedThresholdImageFilter
>>>>>>
>>>>>
>>>>>
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1ConnectedThresholdImageFilter.html
>
>>>>>>2) Set seed points inside the Trackea
>>>>>> (one seed point should be enough
>>>>>>
>>>>>>3) Set the lower threshold to -700 Hounsfield Units
>>>>>>
>>>>>>4) Set the upper threshodl to -200 Hounsfield Units
>>>>>>
>>>>>>5) Run the filter
>>>>>>
>>>>>>
>>>>>>This shoudld give you a binary mask of the two lungs.
>>>>>>
>>>>>>If you find many holes inside the mask, you could
>>>>>>clean it up with the Morphological filters or the
>>>>>>Voting Hole filling filters.
>>>>>>
>>>>>>
>>>>>> Start typing.... :-)
>>>>>>
>>>>>>
>>>>>>Once you are done, please keep in mind that we are
>>>>>>happy to get "case studies" as papers to the Insight
>>>>>>Journal. http://www.insight-journal.org/
>>>>>>
>>>>>>These, more than "papers", are actually "technical
>>>>>>reports". Yours could be entitled "Using ITK for
>>>>>>Lung segmentation".
>>>>>>
>>>>>>Please let us know if you need any help on writing
>>>>>>this report.
>>>>>>
>>>>>>
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>>
>>>>>> Luis
>>>>>>
>>>>>>
>>>>>>
>>>>>>------------
>>>>>>wassim_belhadj at topnet.tn wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>>Hi,
>>>>>>>
>>>>>>>I'am using VTK and Qt4 for volume rendering.
>>>>>>>
>>>>>>>I'm trying to define values of pixel that define LUNGS.
>>>>>>>
>>>>>>>I assigned opacity to each pixel by checking its intensity value
>
> with
>
>>>>>>>Hounsfield Units values.
>>>>>>>
>>>>>>>I obtained a volume that contains the lungs and skin/air.
>>>>>>>
>>>>>>>
>>>>>>>I varied the opacity values to eliminate the skin that surrounds the
>>>>>>>lungs
>>>>>>>but I have not succeeded.
>>>>>>>
>>>>>>>VTK users asked me to do a segmentation (region growing algorithm )
>
> to
>
>>>>>>>display ONLY the lungs.
>>>>>>>
>>>>>>>http://www.vtk.org/pipermail/vtkusers/2009-February/099524.html
>>>>>>>
>>>>>>>Can you help me?
>>>>>>>
>>>>>>>How do I proceed (ITK-VTK)?
>>>>>>>
>>>>>>>Have you an example of 3D segmentation (region growing algorithm)?
>>>>>>>
>>>>>>>Regards, Wassim
>>>>>>>_____________________________________
>>>>>>>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
>>>>>>>
>>>>>
>>>>>
>
More information about the Insight-users
mailing list