[Insight-users] Current state of "Active Contours without edges" in Review folder? - Example?
Kishore Mosaliganti
kishoreraom at gmail.com
Tue Jul 28 08:57:07 EDT 2009
Dear Amardeep,
Please take a look at the Insight Journal publication:
http://www.insight-journal.org/browse/publication/642
All the parameters that you mentioned below are completely explained in the
context of the level-set PDE that was implemented. If you want to know more
about the mathematics , the Insight paper further references the following
papers:
1. T. Chan and L. Vese. An active contour model without edges. In
Scale-Space Theories in Computer Vision, pages 141–151, 1999.
2. A. Dufour, V. Shinin, S. Tajbakhsh, N. Guillen-Aghion, J. C. Olivo-Marin,
and C. Zimmer. Segmenting and tracking fluorescent cells in dynamic 3-d
microscopy with coupled active surfaces. Image Processing, IEEE Transactions
on, 14(9):1396–1410, 2005.
Finally, the parameter m_ReinitializationWeight which was added recently is
by incorporating the method explained in the paper:
3. Chunming Li, Chenyang Xu, Changfeng Gui, and Martin D. Fox, "Level Set
Evolution Without Re-initialization: A New Variational Formulation",* CVPR
2005*.
In Insight/Examples/Review, we have placed fully commented example codes for
using the filters -- both dense and sparse for the single phase cases.
Here, I explain the parameters once again:
1. CurvatureWeight - Penalize high curvatures of level set curve or
surface. This also means that if your object to be segmented has small
features on its boundaries, then they will be missed.
2. AreaWeight - Hastens the movement of level-set to its final convergence.
If you initialized in a mostly interior region, then this parameter must be
negative to force the level-set to grow quickly and vice-versa.
3. ReinitializationWeight - This is based on reference 3 above. It forces
the level-set function to remain a signed distance function as it evolves.
You can turn it off (set to 0) in the sparse-field methods since the sparse
implementation is not based on a signed distance function. Only needs to be
used in dense implementation.
4. VolumeWeight - In tracking applications, this will weight the volume term
for the object being tracked.
5. Volume - The volume of the object being tracked in terms of pixels.
Kishore
Finally, please keep checking the CVS version of the code regularly for
getting the latest updates.
Kishore
On Tue, Jul 28, 2009 at 5:40 AM, Amardeep Singh <amar.singh at gmx.de> wrote:
> Hello everyone
>
> Unfortunately, I have not yet received an answer to my question on how to
> interpret the different parameters of
> the "Active Contours without edges" algorithm. Has anyone already used the
> filter and can share his understanding of
> the meaning of the parameters, please?
> Thanks a lot!
>
> Best regards
> Amardeep
>
> Amardeep Singh wrote:
>
>> Dear Kishore
>>
>> Thanks a lot for the bugfix. I can get the program to run with the newest
>> fix.
>> However, I would be interested in the parameters that you used in order to
>> produce the result.
>> Could you give me a little more insight in the meaning of some of the
>> parameters, as well,please?
>> Some are quite clear, but I couldn't find a correspondence for the others
>> in the paper:
>> 1. CurvatureWeight: Penalize high curvature of level set?; where is this
>> in the paper?
>> 2. AreaWeight: Penalize high area of level set surface
>> 3. ReinitializationWeight, in the code also called
>> "SetReinitializationSmoothingWeight":
>> What does this mean? Is it in the paper?
>> 4. VolumeWeight, in the code also called VolumeMatchingWeight:
>> What does it mean? Is it in the paper?
>> 5. Volume: I guess that this the parameter of the volume parametrization,
>> so it penalizes large volumes
>> within the level set surface.
>>
>> Thank you very much!
>>
>> Best regards
>> Amardeep
>>
>> Kishore Mosaliganti wrote:
>>
>>> Hi Amardeep,
>>>
>>> Thanks for the format conversion and also for putting up your data. I was
>>> able to reproduce the problem at my end. Your images had a negative origin
>>> and revealed a bug in code. This has been corrected now in the latest CVS
>>> repository.
>>>
>>> Meanwhile, I have placed an example script and its output:
>>> http://bmi.osu.edu/~kishore/Amardeep/<http://bmi.osu.edu/%7Ekishore/Amardeep/><
>>> http://bmi.osu.edu/%7Ekishore/Amardeep/>
>>>
>>> Please try it out at your end and let me know how things work out.
>>>
>>> Kishore
>>>
>>> On Wed, Jul 22, 2009 at 5:41 AM, Amardeep Singh <amar.singh at gmx.de<mailto:
>>> amar.singh at gmx.de>> wrote:
>>>
>>> Dear Kishore
>>>
>>> I have converted the images to *.mhd, just like Alexander
>>> suggested. Does this solve the issue of
>>> opening the images for you?
>>> Please let me know if you still cannot use them.
>>> Thanks a lot for your help!
>>>
>>> Best regards
>>> Amardeep
>>>
>>>
>>> Oleksandr Dzyubak wrote:
>>>
>>> Hi Kishore,
>>>
>>> You probably have some box with ITK installed.
>>> If you do, you could compile examples from ITK which
>>> comes in separate package "InsightApplications" (if you have
>>> not done so yet).
>>> Once successfully compiled, under directory
>>> ~/InsightApplications/ConvertBetweenFileFormats you can find
>>> the executable
>>> ConvertBetweenFileFormats.
>>>
>>> The rest is quite simple. If you want to use Paraview, do as below
>>>
>>> ./ConvertBetweenFileFormats image.nii image.mhd
>>>
>>> The second operand could be change to whatever your favorite
>>> format is.
>>>
>>> Alex
>>>
>>> Kishore Mosaliganti wrote:
>>>
>>> Hi Amardeep,
>>>
>>> I am unable to open your images correctly or something is
>>> wrong.. What I saw was that the images was having 0s
>>> throughout.
>>>
>>> Can you write them in some other format that I can open in
>>> Paraview, for example?
>>>
>>> Thanks,
>>> Kishore
>>>
>>> On Tue, Jul 21, 2009 at 2:06 PM, Amardeep Singh
>>> <amar.singh at gmx.de <mailto:amar.singh at gmx.de>
>>> <mailto:amar.singh at gmx.de <mailto:amar.singh at gmx.de>>> wrote:
>>>
>>> Dear Kishore
>>> Dear ITK Users
>>>
>>> I have now been trying to use the filter, but,
>>> unfortunately, I
>>> cannot get it to run. My initial level set and my
>>> feature image
>>> are overlapping perfectly (same orientation/position
>>> information
>>> in the headers), but when I run the example program
>>> "ScalarSinglePhase3DTest.cc", I get the following error
>>> (it is
>>> indeed printed out twice):
>>> Exception caught !
>>>
>>> itk::ExceptionObject (0x8152b00)
>>> Location: "virtual void
>>> itk::DataObject::PropagateRequestedRegion()"
>>> File:
>>>
>>> /asingh/workspace/InsightToolkit-3.14.0/Code/Common/itkDataObject.cxx
>>> Line: 397
>>> Description: Requested region is (at least partially)
>>> outside the
>>> largest possible region.
>>>
>>>
>>> Exception caught !
>>>
>>> itk::ExceptionObject (0x8149d88)
>>> Location: "virtual void
>>> itk::DataObject::PropagateRequestedRegion()"
>>> File:
>>>
>>> /asingh/workspace/InsightToolkit-3.14.0/Code/Common/itkDataObject.cxx
>>> Line: 397
>>> Description: Requested region is (at least partially)
>>> outside the
>>> largest possible region.
>>>
>>> My parameters are chosen arbitrarily:
>>> curvature weight: 0.1
>>> area weight: 0.1
>>> laplacian weight: 0.1
>>> volume weight: 0.1 (what does this mean?)
>>> volume: 0 (what does this mean?)
>>> overlap weight: 0
>>>
>>> For your convenience, I have made the images available
>>> under the
>>> following URL:
>>> http://www.doc.ic.ac.uk/~asingh/<http://www.doc.ic.ac.uk/%7Easingh/>
>>> <http://www.doc.ic.ac.uk/%7Easingh/>
>>> <http://www.doc.ic.ac.uk/%7Easingh/>
>>>
>>> I have also put my code there.
>>>
>>> Does anyone have an idea, why it fails?
>>>
>>> Any comment is highly appreciated!
>>>
>>>
>>> Best regards
>>> Amardeep
>>>
>>>
>>> Kishore Mosaliganti wrote:
>>>
>>> Hi Amardeep,
>>>
>>> The names of the classes were changed when moving
>>> to the
>>> Review directory. Therefore, you will need to
>>> modify the main
>>> cxx that you downloaded from Insight Journal.
>>>
>>> Thanks,
>>> Kishore
>>>
>>> On Thu, Jul 16, 2009 at 11:17 AM, Amardeep Singh
>>> <amar.singh at gmx.de <mailto:amar.singh at gmx.de>
>>> <mailto:amar.singh at gmx.de <mailto:amar.singh at gmx.de>>
>>> <mailto:amar.singh at gmx.de
>>> <mailto:amar.singh at gmx.de> <mailto:amar.singh at gmx.de
>>> <mailto:amar.singh at gmx.de>>>> wrote:
>>>
>>> Dear ITK Users
>>>
>>> I would like to use the "Active Contours without
>>> edges"
>>> algorithm,
>>> which is available in
>>> the Review directory if I understood correctly.
>>> So, I am wondering where I can find an example
>>> about how to use
>>> the algorithm. I tried to run the
>>> "ScalarSinglePhase3DTest" example of the Insight
>>> Journal
>>> submission, but it failed to compile.
>>> The first error read:
>>> /applications/ScalarSinglePhase3DTest.cc:4:52:
>>> error:
>>> itkSparseMultiphaseLevelSetImageFilter.h: No
>>> such file or
>>> directory
>>>
>>> I am a little confused by the multitudes of
>>> implementations
>>> that
>>> are available with respect to this
>>> algorithm, at the moment (3 Insight Journal
>>> submission +
>>> Review).
>>> Can someone tell me where to start and how to
>>> get it running?
>>> Thanks a lot!
>>>
>>> Best regards
>>> Amardeep
>>>
>>>
>>>
>>> _____________________________________
>>> Powered by www.kitware.com
>>> <http://www.kitware.com> <http://www.kitware.com>
>>> <http://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
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _____________________________________
>>> Powered by www.kitware.com <http://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
>>>
>>>
>>> _____________________________________
>> 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/20090728/46c57686/attachment-0001.htm>
More information about the Insight-users
mailing list