[Insight-users] ConfidenceFilter

[Samuel Rodríguez Bescos]

Hello!!

I'm using in my own program.

The seed is given by the user. I've used the same way of the RegionGrowing
example. The same way of the homogeneous image in this example.

The input image of the confidence filter is the original image before have
applied an anisotropic imageDiffusion (with 15 iterations, 0.125 as time
Step and 0.5 as constant diffusion ) and a gradientMagnitudeFilter. I put
the seed in the middle of the liver (more or less). And the filter works
with the outputimage of the gradientMagnitudeFilter.

I think that only difference between the example and my program is that I
use the GradientMagnitude image as the input image, but I think that that
can't be the problem, can it?

If there is something you want to Know, tell it to me.

Thnks in advance.

Sam

----- Original Message -----
From: "Luis Ibanez" <luis.ibanez@kitware.com>
To: "Samuel Rodríguez Bescos" <srodrigu@gbt.tfo.upm.es>
Cc: <insight-users@public.kitware.com>
Sent: Tuesday, January 28, 2003 2:46 PM
Subject: Re: [Insight-users] ConfidenceFilter


>
>
> Hi Samuel,
>
> That sounds like a strange behavior.
>
> The filter should converge when you increase
> the number of iterations.
>
> There is however the risk of it growing until
> filling up the entire image, but this is only
> possible if your image is some sort of continuous
> ramp. In this case, each time the mean and variance
> are computed in an iteration, their values will shift
> allowing more pixels to be added to the region.
>
>
> I will rather suspect of your seed point...
> If the seed is planted in a non-typical region of
> the image the filter may end up stuck in a small
> region.
>
> Could you give us more details on the code
> you have around the filter ?
>
> Are you using the examples in
>
>     Insight/Examples ?
>
> or the GUI application in
>
>     Insight/Applications/RegionGrowingSegmentation  ?
>
> Is this your own application ?
> How are you selecting the seed point ?
>
>
> Please let us know,
>
>
>    Thanks
>
>
>     Luis
>
>
> --------------------------------
>
>
>
> Samuel Rodríguez Bescos wrote:
>
> > Hi!!
> >
> >
> >
> > I think that the confidence filter has got a strange behaviour. I set
> > the next the parameters : Initial Radious=1, Multipler=1.0 and
> > iterations 1 and the filter works well. But when I increase the
> > iterations, the filter doesn't  select any zone of the image.
> >
> >
> >
> > is that could be possible?.
> >
> >
> >
> > Thnks!!
> >
> >
> >
> > Sam
> >
>
>
>