[ITK-users] Combining several segmenations

[Felix Burk]

Hello,

I'm using ITKs ScalarChanAndVeseDenseLevelSetImageFilter to segment parts
of a medical image file series. The problem is that the resolution is
pretty low. So far its working well, but I'd like to have more accurate
results.

The images in the file series are similar, but change more after each time
step. Some segmentations work really well but others differ quite a bit
from the desired result.
My idea is to combine segmentations from different time steps, like the
current and previous one. I think this might lead to better results,
because the segmentations should be similar too.
I'm quite new to ITK and image analysis in general, maybe that's why I
couldn't find something useful for this kind of problem.

I already tried to compare them using SquaredDifferenceImageFilter and
SimilarityIndexImageFilter, but I don't know if the results are of any use.
Is there any way to combine segmentations from several time steps to obtain
better results for a single time step?

I implemented the algorithm as a Paraview Plugin, if that matters. I could
provide some Screenshots too if this might help.

Thanks,
Felix

[Matt McCormick]

Hello Felix,

Once approach is to use the LabelVotingImageFilter:

https://itk.org/Doxygen/html/classitk_1_1LabelVotingImageFilter.html

Hope this helps,
Matt

Post by Felix Burk
Hello,
I'm using ITKs ScalarChanAndVeseDenseLevelSetImageFilter to segment parts of
a medical image file series. The problem is that the resolution is pretty
low. So far its working well, but I'd like to have more accurate results.
The images in the file series are similar, but change more after each time
step. Some segmentations work really well but others differ quite a bit from
the desired result.
My idea is to combine segmentations from different time steps, like the
current and previous one. I think this might lead to better results, because
the segmentations should be similar too.
I'm quite new to ITK and image analysis in general, maybe that's why I
couldn't find something useful for this kind of problem.
I already tried to compare them using SquaredDifferenceImageFilter and
SimilarityIndexImageFilter, but I don't know if the results are of any use.
Is there any way to combine segmentations from several time steps to obtain
better results for a single time step?
I implemented the algorithm as a Paraview Plugin, if that matters. I could
provide some Screenshots too if this might help.
Thanks,
Felix
_____________________________________
Powered by www.kitware.com
Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html
http://www.kitware.com/products/protraining.php
http://www.itk.org/Wiki/ITK_FAQ
http://public.kitware.com/mailman/listinfo/insight-users

_____________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html

Kitware offers ITK Training Courses, for more information visit:
http://www.kitware.com/products/protraining.php

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://public.kitware.com/mailman/listinfo/insight-users