3D reconstruction of sections

[Xiaowei Li]

Hi ImageJ users and developers,

     We are trying to reconstruct the 3D structure of a mice organ from
about 200 sections of about 1 um thickness. The sections were H&E stained
and imaged in bright field with 40x objective.  The obtained images were
stitched in order to get full images of all sections

  The images were imported to Fiji as a stack and viewed in 3D viewer.
However, we are not happy about the presentation of the 3D structure. The
cells are filled with colors and are very difficult to see a real 3D
structure. We would like to know whether there is other plugins or package
that can make the reconstructed 3D look like a real 3D.

Thank you in advance.

Best Regards,

Xiaowei





--
Xiaowei Li, Ph.D
Associate Professor
Shanghai Centre for Systems Biomedicine
Shanghai Jiao Tong University
Office:86 21-34206632
Fax:  [phone]

[Joel Sheffield]

One trick that I have used is to invert the colors so that the nuclei become
bright objects, and then use the old 3D image routine in the stacks menu of
ImageJ to make a series of maximum projection images that rotate.  Then you
can invert the colors back and still see the stereo effect.

You might also consider using a subset of the images rather than all 200.

Best of luck.

Joel


On Wed, Jan 12, 2011 at 5:31 AM, Xiaowei Li <[hidden email]> wrote:

> Hi ImageJ users and developers,
>
>     We are trying to reconstruct the 3D structure of a mice organ from
> about 200 sections of about 1 um thickness. The sections were H&E stained
> and imaged in bright field with 40x objective.  The obtained images were
> stitched in order to get full images of all sections
>
>  The images were imported to Fiji as a stack and viewed in 3D viewer.
> However, we are not happy about the presentation of the 3D structure. The
> cells are filled with colors and are very difficult to see a real 3D
> structure. We would like to know whether there is other plugins or package
> that can make the reconstructed 3D look like a real 3D.
>
> Thank you in advance.
>
> Best Regards,
>
> Xiaowei
>
>
>
>
>
> --
> Xiaowei Li, Ph.D
> Associate Professor
> Shanghai Centre for Systems Biomedicine
> Shanghai Jiao Tong University
> Office:86 21-34206632
> Fax:  [phone]
>

--


Joel B. Sheffield, Ph.D
Department of Biology
Temple University
Philadelphia, PA 19122
Voice: 215 204 8839
e-mail: [hidden email]
URL:  http://astro.temple.edu/~jbs

[Albert Cardona-2]

2011/1/12 Xiaowei Li <[hidden email]>:

> Hi ImageJ users and developers,
>
>     We are trying to reconstruct the 3D structure of a mice organ from
> about 200 sections of about 1 um thickness. The sections were H&E stained
> and imaged in bright field with 40x objective.  The obtained images were
> stitched in order to get full images of all sections
>
>  The images were imported to Fiji as a stack and viewed in 3D viewer.
> However, we are not happy about the presentation of the 3D structure. The
> cells are filled with colors and are very difficult to see a real 3D
> structure. We would like to know whether there is other plugins or package
> that can make the reconstructed 3D look like a real 3D.

Dear Xiaowei,

It would help if you described "real 3D".

Have you tried selecting the object in the 3D Viewer, and then going
to "Edit - Display as" and choosing Surface or Orthoslice?

Have you tried changing the transparency, or the threshold for a 3D
surface? In "Edit - Attributes".

Could you make the data available, even if in scaled down form, so
that we can see what the issue is? Could you attach a screenshot of
the 3D image that doesn't look like you expect, with a description of
what is wrong with it?

Albert

--
http://albert.rierol.net

[dscho]

Hi,

On Wed, 12 Jan 2011, Albert Cardona wrote:

> 2011/1/12 Xiaowei Li <[hidden email]>:
>
> >   We are trying to reconstruct the 3D structure of a mice organ from
> > about 200 sections of about 1 um thickness. The sections were H&E
> > stained and imaged in bright field with 40x objective.  The obtained
> > images were stitched in order to get full images of all sections
> >
> >  The images were imported to Fiji as a stack and viewed in 3D viewer.
> > However, we are not happy about the presentation of the 3D structure.
> > The cells are filled with colors and are very difficult to see a real
> > 3D structure. We would like to know whether there is other plugins or
> > package that can make the reconstructed 3D look like a real 3D.
>
> It would help if you described "real 3D".

I agree.

> Have you tried selecting the object in the 3D Viewer, and then going to
> "Edit - Display as" and choosing Surface or Orthoslice?
>
> Have you tried changing the transparency, or the threshold for a 3D
> surface? In "Edit - Attributes".

Also, there is a function in Edit>Attributes called "Transfer function"
where you can modify the transparency (this is called "alpha" as is common
in image processing).

> Could you make the data available, even if in scaled down form, so that
> we can see what the issue is? Could you attach a screenshot of the 3D
> image that doesn't look like you expect, with a description of what is
> wrong with it?

I agree that it is very hard to give you a meaningful feedback if you do
not make example data available.

Ciao,
Johannes