June 1, 2004

Contents

Boron Stack

This exercise loads a stack as a single tiff file and examines it using some of the stack tools.

This is a stack of 50 images. The images were taken by Greg Gillen (NIST) using an Camec 3F Ion microscope. The images are about 150 micrometers across. The sample is polished steel, the images are of Boron that has migrated to the grain boundaries. The ion microscope slowly erodes the sample away as images are taken, so that these images amount to serial sections of the sample. The sections are about half a micrometer apart - the distance scale is not the same as the pixel to pixel distance across any of the images.

 

Image

ImageJ

This looks like just one image. In fact it represents 50 images. Only the first one is displayed, denoted by the (1/50) in the title bar).

A smoother 'movie' results from showing the images in order, then in reverse order, and so on.

Options -> Preferences... menu. Check the Oscillating Movies box, then re-invoke the Stacks -> Animate menu.

 

 

ImageJ:  Image / Stacks / Animation Options

Speed up the movie using 15 frames per second (the default is 7), check the "Loop Back and Forth" option, and ckeck the "Start Animation" option if the movie is not running already.

 

Another way to view all of the images at once is to make a montage.

 Default parameter settings. Settings used here.  ImageJ Montage options used here

  (Note - a montage of any images of the same size can be made in this way. First use the Stacks -> Windows to Stack Image / Stacks / Convert Images to Stack menu, then the Image / Stacks -> Make Montage... menu.)

Image

ImageJ

 

 

The stack can be viewd from the side (if you consider the images here as from the top) by reslicing:

This looks a little 'thin'. To make the stack behave as if it were thicker:

This result could be obtained with the thin image above:

This single image result is identical to the 198-10.43 window above (names differ in ImageJ).


3-D Projection

The 'very cool', but time and memory intensive way to see a 3-D view of the grain boundaries is to project the stack. With ImageJ, the computation is fast.  With the newer computers, memory does not seem to be an issue.

Setting the slice spacing to 3 is necessary for the Project... menu to give the more realistic result. (The older version of Image using the Stacks macro achieved the equivalent of Slice spacing = 3 by replicating slides, which in this case, tripled the memory required to hold the stack. With the current version of Image, I don't believe there is any disadvantage of having a slice spacing greater then 1, except for viewing the stack exactly edge-on - see see below.).

 

 

Image stack projection options:

Image / Stacks / 3D Projection options

The only changes from the default settings are "Brightest Point", and Slice Spacing = 3 pixels.

The result is another stack, which can be viewed as a movie (turn off oscillating movies) to give a realistic impression of the 3D grain boundary. This 'movie' is as if each of the 50 original images were printed on clear plastic, physically stacked one above the other, and rotated.

To rotate the stack


(Adaption to ImageJ not done below.)

Here is a montage of the stack:

Settings:

=

Montage:

 

Slice Replication

If the slices are replicated, the edge-on view looks more like the other views. This exercise may exceed the memory of your machine:

This triples the number of images in the stack, but now there are no 'gaps'. After projection as above, the edge-on (a) and just after edge-on (b) views in each case look like this:

 

 

 Slice spacing x3

 Replicated Slices x 3

 a    
 b    

With stack replication, the 'jump' when going past the edge on view is less than with the original stack and x3 slice spacing. The edge view (90 deg, 270 deg) can be avoided by altering the parameters for incrementing the angle when projecting.


 

Stereo

The illusion of three-dimensions is also obtained by just rocking between two slices of the projection stack.

Rather than rocking, two adjacent slices may be viewed in stereo. In this case, they must be rotated 90 deg. so that the apparent axis of rotation is vertical.

First, we need to get two slices (say, 1/36 and 2/36) into separate windows. There is a Stack to Windows menu, but that will make 36 windows - we wish to isolate only two slices.

The new windows, each with one of the projection slices, look like this (half size):

 

The images may now be viewed in stereo. The hole in the middle (or apex) is on the top, or closest to you.

Cross-Eyed

 Right

 Left

Bug - Eyed

 Left

 Right


Stereo (Red / Blue) Analyglyph

To make a red-blue anaglyph,