[Insight-users] Recommendation of algorithms for liver vessel extraction

[Juan Cardelino]

Dear Xiaopeng,
                   I think most of us are not familiar with liver
structure, and how it shows in CT. Maybe providing us with and example
would help us advice you about the algorithms to use.
In addition, If I were you, I would try to design and answer a set of
questions like the following:
*which features define the objects to be segmented?
** have vessels a constant color inside and different from the background?
** have they contrasted boundaries?
** would it be possible to identify a texture or regular pattern
inside and/or outside ?
*how many types of objects do you need to detect?
**here you seem to have 3 objects and the background
* how fast doest the algorithm need to be?
*how robust the algorithm need to be?
**which type of variation do you see when looking at different cases?

And I'm sure there are other useful questions to make. But the point
is, try to describe the images, the objects to be detected and the
restrictions in your problem, and the number of algorithms satisfying
that requirements will narrow down to a few options.
In addition, when asking in a list like these, the more information
you provide, the highest probability of getting useful help from it.
Best regards,
                    Juan


On Mon, Jun 21, 2010 at 2:05 AM, Xiaopeng Yang <yxp233 at postech.ac.kr> wrote:
> Dear ITK users,
>
> I tried to apply region growing algorithm (connected threshold) to extract
> portal vein, hepatic vein, and artery in liver from CT image. However, since
> the intensity values of those three vessels are similar to each other, the
> connected threshold algorithm cannot separately extract those three vessels,
> furthermore, other organs are also extracted.
>
> Does anyone of you have good ideas about this problem? Are there other
> better segmentation algorithms to extract the three vessels separately?
>
> Thank you for your help!
>
> Best,
> Xiaopeng
>
>
>
>
>
> _____________________________________
> 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.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
>