Copy volume from within stack to new stack

[Michael Doube]

Benjamin Schmid wrote:
> Hi Mike,
>
>> Are you posting new versions of the ImageJ 3D viewer to the neurofly page
>> http://132.187.25.13/home/?category=Download&page=Viewer3D
>
> Yes, have you given it a try? Did it work?
>

Yes and mostly.  Easiest was to open a directory of stacks rather than a
4D hypervolume. If I concatenate the stacks then convert to hyperstack
(Image->Hyperstack->Stack to Hyperstack) then open 3D Viewer I just get
the 3D stack displayed with no 4D functions.

Nice work, thanks.

Mike

[Michael Doube]

> Yes and mostly.  Easiest was to open a directory of stacks rather than a
> 4D hypervolume. If I concatenate the stacks then convert to hyperstack
> (Image->Hyperstack->Stack to Hyperstack) then open 3D Viewer I just get
> the 3D stack displayed with no 4D functions.

Ack.  Helps if I use the computer with the updated 3D Viewer installed.
  Now it works fine on a 4D hyperstack (although the hyperstack doesn't
have the same voxel size as the original stacks, so the z-dimension is
squashed).

[dscho]

Hi,

On Fri, 23 Jan 2009, Michael Doube wrote:

> the hyperstack doesn't have the same voxel size as the original stacks,
> so the z-dimension is squashed.

Is this a problem in the conversion to hyperstack (I just looked; it
should not touch the voxel size, except when a composite image is created,
in which case the voxel size is copied), or in the 3D viewer?

I am a bit puzzled, as my experiments with the 3D viewer respected the
voxel size quite well (e.g. if you open the MRI Stack sample, you have to
adjust the voxel dimension so that the head does not appear flattened).

Ciao,
Dscho

[Michael Doube]

Johannes Schindelin wrote:

> Hi,
>
> On Fri, 23 Jan 2009, Michael Doube wrote:
>
>> the hyperstack doesn't have the same voxel size as the original stacks,
>> so the z-dimension is squashed.
>
> Is this a problem in the conversion to hyperstack (I just looked; it
> should not touch the voxel size, except when a composite image is created,
> in which case the voxel size is copied), or in the 3D viewer?
>
> I am a bit puzzled, as my experiments with the 3D viewer respected the
> voxel size quite well (e.g. if you open the MRI Stack sample, you have to
> adjust the voxel dimension so that the head does not appear flattened).

My data is 512 x 512 pixel 16-bit CT scans with DICOM header.  Stack
length varies (~100-200 slices), and slice thickness varies (1-5mm).
   I've standardised stack length by deleting slices.  If I concatenate
(Plugins->Stacks->Concatenate) then convert to Hyperstack
(Image->Hyperstacks->Stack to Hyperstack...), I get a hyperstack.

If I run 3D Viewer on the open hyperstack, it is opened as an ordinary
3D stack, which is not the expected behaviour - I expect 4D
representation in the 3D Viewer.

However, I can save the hyperstack as tif and open it from the 3D Viewer
(File->Load 4D data), where I can play the sequence of 3D renderings
over time.  Neat!

In creating the hyperstack, if I concatenate stacks with different voxel
sizes, the whole concatenated stack is assigned the voxel size of the
first stack.  If I convert and save as a hyperstack then re-open,
Image->Properties tells me the assigned voxel size, but the 3D Viewer
seems to ignore it and gives voxels a 1x1x1 dimension.  I will play with
  this some more and see if I can work out what's going on.  I guess
this is a weird use case - when I look at series of microCT stacks with
uncalibrated, isometric voxels, everything works fine.

A small request for Benjamin:  Could you please put in a dissection
plane (or planes) that can be moved towards and away from the viewing
plane, so that voxels between this plane and the viewing plane are not
rendered? In this way we could see inside our 3D structures,
nondestructively.

Mike