[Insight-users] Re: ITK PROJECT : Vertebra segmentation from MRI

[Luis Ibanez]

Hi Jordi,


Welcome to ITK,

Your project sounds quite interesting.

There is a good number of segmentation algorithms
available in ITK that can help you in your task
of extracting a vertebra from MR images.

Note however that the first thing you must ensure
is that your MR image have enough resolution between
slices.

It is unfortunately too common to acquire MRI datasets
where the inter-slice spacing is 3mm or more. Given that
a typical vertebrae is about 25mm to 30mm in height, such
datasets with large spacings are useless for shape extraction.

The first thing that you should do when you receive a dataset
is not to look at the slices but to the orthogonal view to them.

In this way you will appreciate how much the inter-slice spacing
may complicate the shape extraction, and you will also be able
to asses how much acquisition noise is present in the slices.

In order to get familiar with ITK you *MUST* read the ITK
Software Guide.


    http://www.itk.org/ItkSoftwareGuide.pdf

In particular the "Segmentation" Chapter. Where you will find
details on how to use region growing methods, Level Sets,
Watersheds as well as deformable models.



Please let us know if you have further questions,



    Thanks



      Luis



----------------------------
jordi targa (salle) wrote:
> Hello Luis,
> 
>  
> 
> Sorry, my English isn’t good. I hope you don’t have problems to read it.
> 
>  
> 
> My name is Jordi, I’m a doing my project in “Enginyeria La Salle”, I’m 
> ending Multimedia Engineery.
> 
>  
> 
> I would like participate in a project for ITK and I would like it was my 
> project to end my degree.
> 
> Can you help me?
> 
>  
> 
> In my first step I wanted rebuild a vertebra from RM, My idea was align 
> different series and
> 
> create and algorithm  like FloodFill and propagate it in the series.
> 
> We have several series in different directions.
> 
>  
> 
> Ok, the next step wills the body. I would like use marching cubes and we 
> will have our vertebrae.
> 
>  
> 
> More or less... it’s the typical project. I’m open to do something new. 
> I would like receive your help, an idea…
> 
>  
> 
> I’m not alone in this project and the school has experience in this things.
> 
> I would like introduce ITK in my university and the do more projects 
> together.
> 
>  
> 
> And idea, I would implement for ITK and my university a thing like snake 
> 3D. I’m sure you know snake3D.
> 
> It will be nice extend the closed regions in 3D, in the whole directions 
> of the series.
> 
> In the end we will get level curves.
> 
>  
> 
> After it, we could use a sphere and we can deform this sphere using the 
> level curves.
> 
> We will get the surface but we can use the level curves to get 
> information from the interior of the body.
> 
>  
> 
> I don’t know; if you are interest say me something. I try to write more 
> about it.
> 
>  
> 
> You don’t worry I will have help to write it in English :-D
> 
> I understand It have been difficult read the email.
> 
>  
> 
> Thanks.
> 
>  
> 
> Jordi.
> 
>  
> 
> P.S.:My teacher have experience, left ventricle reconstruction was his 
> thesis.
> 
>  
> 
>  
> 
> 
>  
>