[Insight-users] Re : Hessian3DToVesselness parameters

[Oleksandr Dzyubak]

Hi Syrine,

1)
Yes, I do confirm that *.hdr discribes the Mayo Analyze 7.5 image format.
 
http://www.mayo.edu/bir/PDF/ANALYZE75.pdf

http://imaging.mrc-cbu.cam.ac.uk/imaging/FormatAnalyze

2)
Yes, the ITK library does support that format.

3)
Most likely you are talking about the 
"Hessian3DToVesselnessMeasureImageFilter" filter.

http://www.itk.org/Doxygen/html/classitk_1_1Hessian3DToVesselnessMeasureImageFilter.html

That filter is based on early work by Yoshinobu Sato et al.
References:
    "3D Multi-scale line filter for segmentation and visualization of 
curvilinear structures in medical images", Yoshinobu Sato, Shin 
Nakajima, Hideki Atsumi, Thomas Koller, Guido Gerig, Shigeyuki Yoshida, 
Ron Kikinis.

http://www.spl.harvard.edu/pages/spl-pre2007/pages/papers/yoshi

About parameters.

a)
Sigma.
If correctly normalized, the filter with parameter Sigma=D gives maximum 
response
in voxels which have very high similarity to belong to an tubular object 
with diameter d ~ D.

b)
Alpha1.
Measure of similarity to be a tubular object.
 From the original article by Sato et al., Alpha1 has to be 0.5 which is 
defaulted
in the "Hessian3DToVesselnessMeasureImageFilter" filter.

However it creates some problems  in the areas with singularities .
See my discussions on this forum 
"itkHessian3DToVesselnessMeasureImageFilter false aura at the ends".
 
http://www.nabble.com/itkHessian3DToVesselnessMeasureImageFilter-false-aura-at-the-ends-to25016304.html

Luis Ibanez suggested somewhere (you can search this forum yourself) to 
use Alpha1 = 0.1
It helps a bit but does not solve a problem.

b)
Alpha2.
Helps to recover vessels in the areas with inhomogeneous contrast or 
partial  intensity loss.
Sato recomends Alpha2 = 2.0 which is defaulted in the 
"Hessian3DToVesselnessMeasureImageFilter" filter.

Luis Ibanez also suggested to use Alpha2 = 2.0.

In most cases you work with the images of good quality thus I would 
recommend keeping
the default values.

Regards,

Alex

Syrine Sahmim wrote:
>
> Hi,
> You confirm that .hdr format is supported by itk, that's what you mean?
> can you give me more details specially about sigma, alpha1 and alpha 2 
> arguments please?
>
> Regards
> Syrine
>
> ------------------------------------------------------------------------
> *De :* Oleksandr Dzyubak <adzyubak at gmail.com>
> *À :* John Drescher <drescherjm at gmail.com>
> *Cc :* Syrine Sahmim <syrine.sahmim at yahoo.fr>; Insight Users 
> <insight-users at itk.org>
> *Envoyé le :* Jeudi, 20 Août 2009, 22h35mn 42s
> *Objet :* Re: [Insight-users] I still have error
>
> Hi,
>
> That is the Mayo Analyze 7.5 image format which indeed is supported by 
> ITK.
>
> Alex
>
>
> John Drescher wrote:
> > On Thu, Aug 20, 2009 at 11:53 AM, Syrine 
> Sahmim<syrine.sahmim at yahoo.fr <mailto:syrine.sahmim at yahoo.fr>> wrote:
> > 
> >> I do what you said but i obtain another error message. i joined the 
> print
> >> screen
> >>
> >>   
> >
> > Are you sure .hdr files are supported by itk? I am not familiar with
> > that format.
> >
> > John
> > _____________________________________
> > Powered by www.kitware.com <http://www.kitware.com>
> >
> > Visit other Kitware open-source projects at
> > http://www.kitware.com/opensource/opensource.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
> > 
>
>