[ITK] Restricting region size in region growing filter

Lowekamp, Bradley (NIH/NLM/LHC) [C] blowekamp at mail.nih.gov
Wed Jul 26 08:31:04 EDT 2017


Hello,

There are many ways to do this operation and the precise definition of size for the application.

Here are some brief suggestions:

- Combine multiple masks. Such a doing a global threshold, then combining it with a mask which limits the size, then run connected components to choose
- Look into the FastMarching methods, perhaps using a speed image of a global threshold
- Morphological methods such as reconstruction by dilation, with the seeds as the marker image and the global threshold as the maske image.

I hope these suggestions get you started. I would recommend interactively exploring these algorithms with SimpleITK of ITK Python. 

You can look at these Jupyter notebooks for some ideas:
https://github.com/InsightSoftwareConsortium/SimpleITK-Notebooks/tree/master/Python

HTH,
Brad


On 7/26/17, 3:17 AM, "Alexander Bruy" <alexander.bruy at gmail.com> wrote:

    Hi,
    
    I use ConnectedThresholdImageFilter to perform region growing with seed points
    and need to be sure that region size does not exceed some value, e.g. 5m or 8m
    by diameter.
    
    In ConnectedThresholdImageFilter docs I can't find any parameters that allows me
    to limit region size. So I'm wondering if ITK has some tools that can
    help with this?
    For example, some filter which can be applied to resulting image or maybe
    a combination of filters that allow to perform region growing with
    seed points and
    limit region size?
    
    Thanks.
    -- 
    Alexander Bruy
    _______________________________________________
    Community mailing list
    Community at itk.org
    http://public.kitware.com/mailman/listinfo/community
    



More information about the Community mailing list