MATLAB Central Newsreader - 3D using ISOSURFACE

3D using ISOSURFACE

arulvel v — Wed, 13 Feb 2008 03:21:02 -0500

I have constructed a 3D matrix and used ISOSURFACE to view
it as 3D image.

The 3D image i get shows layers of slices and gaps between
slices. Is it possible to smooth the 3D image. ?

Is there any other way to reconstruct the 3D matrix to 3D
image, other than isosurface. ?

Please advice.

thanks

Re: 3D using ISOSURFACE

Vihang Patil — Wed, 13 Feb 2008 08:10:06 -0500

"arulvel v" <varulvel@hotmail.com> wrote in message
<fotnmu$8t8$1@fred.mathworks.com>...
> I have constructed a 3D matrix and used ISOSURFACE to
view
> it as 3D image.
>
> The 3D image i get shows layers of slices and gaps
between
> slices. Is it possible to smooth the 3D image. ?

doc smooth3d

> Is there any other way to reconstruct the 3D matrix to 3D
> image, other than isosurface. ?

Volume rendering found here
http://www.mathworks.com/matlabcentral/fileexchange/loadFil
e.do?objectId=4927&objectType=file

HTH
Vihang

Re: 3D using ISOSURFACE

stefanbanev@yahoo.com — Fri, 15 Feb 2008 17:02:26 -0500

On Feb 12, 7:21=A0pm, "arulvel v" <varul...@hotmail.com> wrote:
> I have constructed a 3D matrix and used ISOSURFACE to view
> it as 3D image.
>
> The 3D image i get shows layers of slices and gaps between
> slices. Is it possible to smooth the 3D image. ?
>
> Is there any other way to reconstruct the 3D matrix to 3D
> image, other than isosurface. ?
>
> Please advice.
>
> thanks

The link in post above definitely may help you to do VR in MathLab but
do not expect to have descent quality iso-surface by means VR under
MathLab. While you stick with cloud like representation (low opacity)
with no gradient lighting it may work Ok. Once you try to apply higher
opacity to have more definitive iso-surfaces it will not work well.
The same true for gradient lighting by means of VR (no meshing). Via
super-sampling you may get better result for tiny bricks let say
64x64x64 resampled to 512x512x512 and it is not very practical from my
point of view. Please correct me if you find way to proof me wrong.

--sb

Re: 3D using ISOSURFACE

Vihang Patil — Fri, 15 Feb 2008 18:08:05 -0500

"stefanbanev@yahoo.com" <stefanbanev@yahoo.com> wrote in
message <8357e680-0524-44a0-ba4e-
a4a2fae049c5@i29g2000prf.googlegroups.com>...
> On Feb 12, 7:21=A0pm, "arulvel v" <varul...@hotmail.com>
wrote:
> > I have constructed a 3D matrix and used ISOSURFACE to
view
> > it as 3D image.
> >
> > The 3D image i get shows layers of slices and gaps
between
> > slices. Is it possible to smooth the 3D image. ?
> >
> > Is there any other way to reconstruct the 3D matrix to
3D
> > image, other than isosurface. ?
> >
> > Please advice.
> >
> > thanks
>
> The link in post above definitely may help you to do VR
in MathLab but
> do not expect to have descent quality iso-surface by
means VR under
> MathLab. While you stick with cloud like representation
(low opacity)
> with no gradient lighting it may work Ok. Once you try
to apply higher
> opacity to have more definitive iso-surfaces it will not
work well.
> The same true for gradient lighting by means of VR (no
meshing). Via
> super-sampling you may get better result for tiny bricks
let say
> 64x64x64 resampled to 512x512x512 and it is not very
practical from my
> point of view. Please correct me if you find way to
proof me wrong.
>
> --sb

Do you have a better VR code than vol3d?, with gradient
lighting?.
Vihang

Re: 3D using ISOSURFACE

stefanbanev@yahoo.com — Fri, 15 Feb 2008 19:00:52 -0500

On Feb 15, 10:08 am, "Vihang Patil" <vihang_pa...@yahoo.com> wrote:
> "stefanba...@yahoo.com" <stefanba...@yahoo.com> wrote in
> message <8357e680-0524-44a0-ba4e-
> a4a2fae04...@i29g2000prf.googlegroups.com>...
>
>
>
>
>
> > On Feb 12, 7:21=A0pm, "arulvel v" <varul...@hotmail.com>
> wrote:
> > > I have constructed a 3D matrix and used ISOSURFACE to
> view
> > > it as 3D image.
>
> > > The 3D image i get shows layers of slices and gaps
> between
> > > slices. Is it possible to smooth the 3D image. ?
>
> > > Is there any other way to reconstruct the 3D matrix to
> 3D
> > > image, other than isosurface. ?
>
> > > Please advice.
>
> > > thanks
>
> > The link in post above definitely may help you to do VR
> in MathLab but
> > do not expect to have descent quality iso-surface by
> means VR under
> > MathLab. While you stick with cloud like representation
> (low opacity)
> > with no gradient lighting it may work Ok. Once you try
> to apply higher
> > opacity to have more definitive iso-surfaces it will not
> work well.
> > The same true for gradient lighting by means of VR (no
> meshing). Via
> > super-sampling you may get better result for tiny bricks
> let say
> > 64x64x64 resampled to 512x512x512 and it is not very
> practical from my
> > point of view. Please correct me if you find way to
> proof me wrong.
>
> > --sb
>
> Do you have a better VR code than vol3d?, with gradient
> lighting?.
> Vihang- Hide quoted text -
>
> - Show quoted text -

vp> Do you have a better VR code than vol3d?, with gradient
vp> lighting?.

Yes, look for "High Definition Volume Rendering" ; I played with it
and may say it is way better then any VR solution out there. I would
love to see it build-in MathLab.

--sb

Re: 3D using ISOSURFACE

arulvel v — Sat, 16 Feb 2008 03:25:06 -0500

thanks for the suggestion.

i have tried with vol3d from the fileexchange. previously i
used to use isosurface command directly with the 3d matrix
that i have generated. with the vol3d it look better than that.

i will look for HD Volume Rendering.

I have the following queries.

I have a 3D matrix of the CT images. Is it possible to
reconstruct them to 2D images at any other angle other than
sagital, coronal. ie., other than horizontal and vertical.

Basically i am trying to reconstruct the colon from the CT
image. I have segmented the colon from the CT image and
constructed a 3D matrix. The colon is folded in our body.
So, when seen(reconstructed parallel to the body) from the
front of the body, the folded colon can be viewed. At this
angle it will be clear to see the colon.

The image will have to be reconstructed perpendicular to the
surface of the colon. Since the colon is folded, it is
neccessary to reconstruct them at oblique angles. Is it
possible?

If the surface/volume rendering result were good, i thought
of using the campos and camva of the matlab to navigate
(user guided) through the colon, by manually getting the
camera position from the user. But the images that i have is
not great, so the 3D is not that good to navigate. So i
thought if i could view the colon as 2D at particular angle
and reconstruct them.

thanks