[Insight-users] Concept for edting segmented mask images

[Dawood Masslawi]

I have worked with the Canny edge detector in the ITK and the results were good.
The Affine transformation supports shearing but not inflation unless you can treat the
inflation as a linear scaling which means that if you need to inflate the contours 
non-linearly this transformation function is probably not the way to do it. In that case 
you can use the B-Spline deformable transformation (easier with a warping model 
rather than setting the B-Spline coefficients) or the kernel transforms (need landmarks). 
I think  using the Affine transform will save you a lot of headache and extra
work, so use that one if it suits your application. 
You can make the transformation interactive by changing the transformation parameters
based on user input or use the kernel transforms which require a set of landmarks. There
are plenty of code samples available in the "examples" directory of the ITK for all of the
aforementioned approaches and of course the software guide has further details,
documentation-wise :)
You can generate a mask image using the ImageMaskSpatialObject, a sample code
is included in the "examples" directory of the ITK (Examples\SpatialObjects\
ImageMaskSpatialObject.cxx).
Let us know if you had anymore problems.
HTH,
Dawood

--- On Tue, 3/8/11, Xiaopeng Yang <yxp233 at postech.ac.kr> wrote:

From: Xiaopeng Yang <yxp233 at postech.ac.kr>
Subject: RE: Concept for edting segmented mask images
To: "'Dawood Masslawi'" <masslawi at gmail.com>
Cc: insight-users at itk.org
Date: Tuesday, March 8, 2011, 11:19 AM

Hi Dawood,  Thanks very much for your help!  Could you suggest one specific edge detection filter in the ITK?  In my case, I need to shear and inflate the contours. Could you suggest which transformation function in the ITK is the best for these two changes? And how to apply the transformation function to shear and inflate the contours? Is it interactive?  And how to generate mask images after resampling?  Sorry for my limited knowledge on these topics. I raised many questions.  Best regards,Yang  发件人: Dawood Masslawi [mailto:masslawi at gmail.com] 
发送时间: 2011년 3월 9일 수요일 오전 12:57
收件人: yxp233 at postech.ac.kr
抄送: insight-users at itk.org
主题: RE: Concept for edting segmented mask images  Hi Xiaopeng,  You can use one of the available edge detection filters in the ITK to extract the boundary   contours. Changing the shape of the contours could be a little bit more tricky, my  suggestion is to adapt one of the transformation functions in the ITK based on the kind of  change you want to induce in the contours (e.g. scaling, shearing, inflation and etc.).   Of course there are numerous other ways as well, but ultimately the best option depends  on the kind of change you want to make.  After resampling you should have no problems generating the mask images.  Best regards,  Dawood    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<      Dear Users,     I utilized itk segmentation algorithm to extract the liver from CT images.The results are a series of binary images (mask
 images). For editing thebinary images, for example deleting some part, I have an idea as follows.     First, generating boundary contours from binary images.  Second, changing the shape of the contours.  Finally, generating mask images from the edited contours.     Could you tell me whether this idea is feasible or not? If not, please giveme better ideas.     Thanks,  Yang  


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20110308/748a6e9c/attachment.htm>