[Insight-users] Re: Marker controlled watershed, how to realize in ITK

[Gaëtan Lehmann]

Hi,

Funny to see a reply to a mail old of 2 years :-)

I just want to give some precisions:

The morphological reconstruction is in ITK already. It's the Robinson's  
algorithm I have adapted to work with any type of pixel, not only 8 bits.  
It sadely cause some efficiency problem compared to the original  
algorithm. The version implemented by Richard is a lot more efficient both  
in memory and execution time :-)

Also note that Richard and I have made some effort since this time, to  
improve the mathematical morphology in ITK. The marker controlled  
watershed discussed in this mail for example, can now be found in the  
insight journal, with some others:

- consolidated morpholgy  
(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D6%26order%3Drating%26revision_display%3Dcombined&pubid=124)
- Improving features and performance of binary erode and dilate filters  
(Improving features and performance of binary erode and dilate filters)
- The watershed transform in ITK - discussion and new developments  
(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D4%26order%3Drating%26revision_display%3Dcombined&pubid=92)
- Improving performance of morphological reconstruction  
(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D4%26order%3Drating%26revision_display%3Dcombined&pubid=76)
- MinimaImpositionImageFilter  
(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D3%26order%3Drating%26revision_display%3Dcombined&pubid=63)
- Finding regional extrema - methods and performance  
(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D3%26order%3Drating%26revision_display%3Dcombined&pubid=65)
- Binary morphological closing and opening image  
filters(http://insight-journal.org/InsightJournalManager/view_reviews.php?back=publications.php%3Fjournalid%3D3%26order%3Drating%26revision_display%3Dcombined&pubid=58)

All those article are waiting to have 3 reviews to be moved to the toolkit  
- please add a review to some of them !

Regards,

Gaetan


Le Fri, 06 Oct 2006 16:16:01 +0200, Josiah Yoder <epics.ta at gmail.com> a  
écrit:

> I just want to comment that the morphological reconstruction filters
> are now available in ITK, as discussed in this ITK paper:
>
> http://insight-journal.org/dspace/handle/1926/169
>
> Improving performance of morphological reconstruction
> Beare, Richard
> 15-Feb-2006
>
> itkReconstructionImageFilter
>
> Josiah Yoder
>
> ----- Original Message ---
> *Miller, James V (Research)* millerjv at crd.ge.com
>  <insight-users%40itk.org?Subject=%5BInsight-users%5D%20Re%3A%20Marker%20controlled%20watershed%2C%20how%20to%20reali%0A%09ze%20in%20ITK&In-Reply-To=>*Wed
> Nov 24 10:48:15 EST 2004*
>
>
> I believe the morphological reconstructions are related to the
> geodesic morphology algorithms.  There are a number of geodesic
> morphology algorithms in ITK
>
> itkGrayscaleConnectedClosingImageFilter.h
> itkGrayscaleConnectedOpeningImageFilter.h
> itkGrayscaleFillholeImageFilter.h
> itkGrayscaleGeodesicDilateImageFilter.h
> itkGrayscaleGeodesicErodeImageFilter.h
> itkGrayscaleGrindPeakImageFilter.h
> itkHConcaveImageFilter.h
> itkHConvexImageFilter.h
> itkHMaximaImageFilter.h
> itkHMinimaImageFilter.h
>
> Jim
>
>
>
> -----Original Message-----
> From: Quan Chen [mailto:quanchen at gmail.com
> <http://www.itk.org/mailman/listinfo/insight-users>]
> Sent: Tuesday, November 23, 2004 6:02 PM
> To: insight-users at itk.org  
> <http://www.itk.org/mailman/listinfo/insight-users>
> Cc: Luis Ibanez
> Subject: [Insight-users] Re: Marker controlled watershed, how to realize
> in ITK
>
>
> Also, I am wondering if itk implemented the "morphological
> reconstruction" algorithm.  The following example mention about it.
> http://www.mathworks.com/products/demos/image/watershed/ipexwatershed.html
>
> There is also some papers like:
>
> Luc Vincent, "Morphological Grayscale Reconstruction in Image
> Analysis: Applications and Efficient Algorithms," IEEE transactions on
> Image Processing, vol 2 no. 2 1993.
>
>
>
> On Tue, 23 Nov 2004 16:26:02 -0600, Quan Chen <quanchen at gmail.com
> <http://www.itk.org/mailman/listinfo/insight-users>> wrote:
>> * I just realized someone has already asked this question this month.
> *>* Luis, I don't know whether you proposed would work very well.   As
> *>* that coffee bean example shows.  The objective is to seperate each
> *>* coffee beans from each other.
> *>* http://cmm.ensmp.fr/~beucher/wtshed.html#examples
> <http://cmm.ensmp.fr/%7Ebeucher/wtshed.html#examples>
> *>*
> *>* I belive it is an important feature for the watershed filter.  Since
> *>* it offers more user intenvention so that it can do segmentation
> *>* better.  If we only rely on the image itself, even with Hierarchical
> *>* segmentation, we would still only got crappy result with watershed
> *>* filter.
> *>



-- 
Gaëtan Lehmann
Biologie du Développement et de la Reproduction
INRA de Jouy-en-Josas (France)
tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
http://voxel.jouy.inra.fr