[Insight-users] Region labelling
Sayan Pathak
spathak at insightful . com
Wed, 28 Aug 2002 11:12:59 -0700
Hi Uday,
I agree with Luis. KLM algorithm is an overkill for sure. It is quite =
computation intensive too. Since you already know the seed. The solution =
suggested by Luis is a better one for sure.
Sayan
> -----Original Message-----
> From: Luis Ibanez [mailto:luis.ibanez@kitware.com]
> Sent: Wednesday, August 28, 2002 8:53 AM
> To: Uday Kurkure
> Cc: insight-users@public.kitware.com
> Subject: Re: [Insight-users] Region labelling
>=20
>=20
>=20
> Hi Uday,
>=20
> The KLMRegionGrowImageFilter may be too sophisticated
> for the problem you are trying to solve. This filter
> is capable of segmenting images based on statistical
> caracteristics of the regions. It will be a good choice
> if you don't have already thresholded your image.
>=20
> In the case you mention it seems that the four regions
> of your image are already well defined and you only
> need to get the connected components. If this is the
> case, you may get a more effective result by using the
> ConnectedThresholdImageFilter.
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1ConnectedT
hresholdImageFilter.html
For this filter you need to specify a seed point inside
the region of your interest and a lower and upper threshold
defining the range of gray values for accepting pixels as
being connected to the seed point.
Please let us know is this works for your case,
Thanks
Luis
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Uday Kurkure wrote:
>=20
> Hi all,
>=20
> I have a thresholded image which has 4 regions out of which I want to=20
> extract only the 1 whose seed point I know in advance. For that =
purpose=20
> I m trying to use KLMRegionGrowImageFilter. Can some one show me a =
very=20
> simple sample code to get the image, set the seed point, apply KLM=20
> filter and get the output? Or is there any other algorithm which I can =
> use for the purpose?
>=20
> Regards,
> Uday.
> _______________________________________________
> Insight-users mailing list
> Insight-users@public.kitware.com
> http://public.kitware.com/mailman/listinfo/insight-users
>=20
_______________________________________________
Insight-users mailing list
Insight-users@public.kitware.com
http://public.kitware.com/mailman/listinfo/insight-users