[Insight-users] Segmentation of CT data
sergio.vera at alma3d.com
Thu Apr 7 04:03:33 EDT 2011
The MultiScaleHessian Family of filters in ITK has a way of detecting
plate-like structures (
so maybe the table can be detected as a high response to that kind of filter
at the adequate scale. Give it a try to see if it works.
The advantage is that no seed points would be required to to this.
2011/4/6 Lodron, Gerald <Gerald.Lodron at joanneum.at>
> I do not want to segment heart, lung etc seperately, i want to segment
> human vs background/patient table, especially for CT head data.
> I don't think that this is a so special problem, in my opinion every time
> you want to register a ct data set you need such a segmentation for
> excluding background so i think this a very common task. I already wrote a
> segmentation algorithm based on the itk region grower with seed points in
> the image center but i get sometimes problems with the patient table (and i
> want it 100% automatic) so before i make complicated algorithms i want to
> first ask the community.
> best regards
> *Von:* insight-users-bounces at itk.org [mailto:
> insight-users-bounces at itk.org] *Im Auftrag von *Jon Haitz Legarreta
> *Gesendet:* Mittwoch, 6. April 2011 10:48
> *An:* insight-users at itk.org
> *Betreff:* Re: [Insight-users] Segmentation of CT data
> Dear Gerald,
> the segmentation problem is highly specific and depends on the tissue you
> wish to segment. If you want to segment every organ/tissue in a CT dataset,
> I would dare to say that you will need to look for the most suitable
> algorithm for each of them.
> The ITK Software Guide presents an interesting range of segmentation
> algorithms that you can use your starting point.
> So there is not a one-for-all, all-purpose algorithm. Segmenting all
> tissues you may have in a certain volume can be hard work.
> That said, I haven't had the opportunity to approach the patient-table
> removal problem. But if your requirements allow for some user interaction,
> you can imagine the user selecting a volumen of interest that excludes the
> table (may be too optimistic), or the user picking some seed points on the
> tissues you want to segment and the developping some of your algorithms
> around these points (this would be a typical starting point for the region
> growing algorithm you used). Granting the user some control when overflows
> occur could also be helpful.
> That way, in principle, you could avoid the problem. But I agree that it
> may not be that easy.
> Best regards,
> JON HAITZ
> El 05/04/2011 15:51, Lodron, Gerald escribió:
> Is there a "state-of-the-art" segmentation filter for robust segmentation
> of the human body (with lung etc. without patient table) in a CT data set in
> ITK? I am pretty sure that I am not the first one which needs that...
> If not, does anyone knows a good algorithm for that. I made one on my own
> with region growing but get sometimes problems with the patient table...
> best regards,
> Powered by www.kitware.com
> Visit other Kitware open-source projects athttp://www.kitware.com/opensource/opensource.html
> Kitware offers ITK Training Courses, for more information visit:http://www.kitware.com/products/protraining.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
> Kitware offers ITK Training Courses, for more information visit:
> Please keep messages on-topic and check the ITK FAQ at:
> Follow this link to subscribe/unsubscribe:
Alma IT Systems
C/ Vilana, 4B, 4º 1ª
T. (+34) 932 380 592
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Insight-users