4/9/98

1) Is there a way to create the red/green 3D images using stereo pair

TEM mages without buying an expensive 3D imaging program?

2) What thickness of sections would be needed for TEM work to get

the effect of 3D depth?

3) What is the best angle? I would be using a Philips 410LS

w/goniometer tilt stage. I'm going to hit the library but tips from the horses

mouth are always much better. Thanks in advance.

Rick Vaughn

RLVAUGHN@MAIL.UNMC.EDU

We collect the left and right B/W images using a 4 to 6 degree tilt between

the views for our SEM work. Much more than 6 degrees gets unnatural. It

represents crossing the eyes too much.

Since I work with the SEM, I don't know what the sample requirements would

be for TEM. I would think as long as you have some topographic relief you

should be okay. Even layers of atoms might be enough if you are at high

enough mag.

Once you have the images, you can easily change the hues for the one from

gray to red and the other from gray to green or blue using programs like

PhotoShop. Preparing the anaglyph simply requires merging the two colored

images into one true-color image. Since one image would supply the red

intensity and the other the green intensity, it should be straightforward -

I can think of how I would write a program to do it. But maybe the Photoshop

gurus can tell what buttons to push to make it happen.

Warren E. Straszheim

23 Town Engineering

Iowa State University

Ames IA, 50011

Phone: 515-294-8187 FAX: 515-294-4563

E-Mail: wesaia@iastate.edu

http://www.marl.iastate.edu

I use Photoshop to do this in my bio TEM class, but you should be able to do

it in NIH Image (can't beat the price). You could theoretically do on any

section thickness, the effect depending on the magnification of the

structure viewed. You make each image a different color, adjust the

transparency and overlay them. You can view them on screen or output

them to various hard copy media.

I usually have the students use their replicas (usually freeze-fractured

yeast) since these have a 3-d topography, but I've also done this with

chloroplasts in "thick" (about 250nm) section (don't stain them if you've

used OsO4); they are just O.K. at 100kV. Depending on what you want to

see, you can leach out a lot of background density by using KMnO4 and

maybe work with thicker sections. More volts should also help as would an

energy filter.

The best angle is dependent on the effect desired. To get an idea of the

range of sizes and angles you can work with you might get a hold of

Heuser, 1989 Journal of Electron Microscopy Technique 13: 244-263 and

Steere, R.L. In Chapter 5 of Current Trends in Morphological Techniques

Vol 2 CRC Press 1981. These deal mostly with conventional stereo pairs

and replicas but the approach and geometry is similar .

cheers,

John Heckman

TEM Supervisor

Center for Electron Optics

Michigan State University>

heckman@pilot.msu.edu

working on just this sort of thing with the SEM. I had fairly good success

using Photoshop. The trick is making sure that your images are alligned

vertically, and playing around with the color until it matches the color

of the filters on the viewerlenses. As far as the tilt angle, that depends

on

the magnification and the degree of roughness of the surface being imaged.

Generally speaking, a smoother sample will require more tilt (7-15

degrees) than a rough one (3-7 degrees). At higher magnification, too much

tilt can mean too much parallax (displacement) and the result may be

uncomfortable to view. Usually 4-10 degrees of tilt will produce

sufficient parallax. If still unsure, you can do what I did and take

a

series of images, each at 3 degree differences in tilt. Just be sure that

when you refocus the image after tilting that you do so without changing

the magnification. On the SEM this means using the Z axis control instead

of the objective lense control. Also be careful when you tilt that you

maintain your field of view (tracing the outline of some major features

with a grease pencil on the monitor screen works great).

I'm not sure of how you're planning on tilting your sample. Because you're

using a TEM you may be best off tilting your beam instead of your sample.

I've never tried this as our instrument wasn't able to do this, but it

always seemed like a great way to go. I imagine there would be less

refocusing problems. Less change in contrast too. Let me know how it

works out if you try this method. There are two pretty good references on

this technique you might want to check out. One is "The Perception and

Measurement of Depth In the SEM", by A.Boyde in SEM 1979 Vol 2 page 67-78

(the part you want starts on pg 70). The other is "Introduction to Stereo

Scanning Electron Microscopy" by Eric Chatfield in Vol 6 of Principles and

Techniques of Scanning Electron Microscopy 1978. There are others I can

give you if you want.

Another method you might try is horizontal displacement instead of

tilting. The trick is to have enough displacement so as to produce

sufficient parallax, but maintain enough similarity in the field of view

that the brain can still fuse the two images. I didn't have much success

with this technique. The above references describe this method as well.

Oh, heres a good one. While working on this little project of mine (which

has now become somewhat of a recurring obsession) I discovered the

National Stereoscopic Association (NSA). These people are very interested

in any form of stereoimaging and would love to hear from you and help you

in any way they can. They even have their own magazine Stereo World which

is filled with you guessed it stereopairs. Its great if somewhat bizarre.

And they have a cool website! Check it out at

nsa-3d.org/nsa-membership.html and have your 3d glasses ready. You might

try contacting Larry at larry@sapphire-star.com to see if he has any

advice or suggestions.

Whew! I didn't realize I had so much to say! I hope this has been of some

help. If you have any other questions or just want to discuss the

frustrations of stereoimaging (and there are a few). Feel free to contact

me at coopera@bigdog.engr.arizona.edu (the world's longest email address).

Good Luck and let me know how it turns out

Anne Marie Cooper

coopera@bigdog.engr.arizona.edu

I have students in my SEM course do an exercise in taking stereo pairs.

It provides for a great show-and-tell at the end of the course. Since optimum

angle depends on surface topography and magnification, I have different

students do different angles and then c;compare end results. We usually use

7o as routine with 14o thrown in. The larger angle usually is overkill. It

is very important to focus with Z-control and try to keep the image centered.

I have done a bit with TEM using a goniometer stage but not enough to

give advise. It does work better if you have a rough surface specimen, either

negative stain or sections in a type of resin that tears as you microtome

(like lowicryls or LR White).

For reconstruction, I realign the images in photoshop, using the center

of the image for alignment purposed. This is a critical step and must be done

accurately for good final results. Using layers makes this routine. I put

the original image (0 angle) in the first layer, cross-hairs on the second

layer to mark my center point, and the second image on the third layer, etc.

Reducing the transparency of the second image lets you easily see the

cross-hairs and first image.

Using color channels, you can change the aligned images into your red and

blue (or green) colors and adjust transparency to overlay them. However, it

is easier to use an image analysis program such as NIH Image (free). I use

ScanAlytic's program, IP Lab Spectrum which makes the conversion about a 20

second process.

The next problem is getting the final image into a form which can be

shown to audiences. I have found that it is often difficult to get the right

exposure using digital slide makers. The slides tend to come out too dark.

You often also have color problems with some films. I have had the best luck

just photographing the computer screen with a 35mm camera. I use Polaroid

Presentation Film which gives you a "what you see is what you get" when

photographing a computer screen. Kodak films are heavy on the blue in this

instance and I haven't tried Fugi film (although it works great in some

digital slide makers!). I use the option in the Adobe PhotoShop tool box of

full screen mode which enlarges the image to maximum screen size and fills

rest of monitor with a black border and then do an exposure series and get

good slides every time.

Have fun with this...your audience will certainly appreciate it.

Debby Sherman, manager

Microscopy Center in Agriculture

Purdue University

emcenter@btny.purdue.edu

Responding to questions 2 & 3 above:

When I got my on-site training after instalation of our Philips CM12, I was

handed a huge compendium of TEM applications tailored for the CM12. There is a

section on stereo-TEM in there which I will copy and send to you by snail mail,

if you'll send me your address.

Within it, is a table and a graph giving suggested tilt angles for stereo TEM as

a funtion of section [or sample] thickness and magnification, which was worked

out long ago by Dr. Lee Peachey. The trend is, the higher the magnification, the

smaller the tilt angle between stereo pairs should be, and the thicker the

sample, the smaller the tilt angle should be.

In response to question #1, so far I haven't played around with the red/green

method of stereo presentation. I make a pair of black and white slides, by

digitizing my TEM negs and use PowerPoint to print them out to a Polaroid slide

maker. Then I use inexpensive stereo slide viewers ($7 ea.,from Reel 3-D

Enterprises, www.stereoscopy.com/reel3d. I have no commercial interest in Reel

3-D.) to view the pairs, which works quite well. I'm presently working out

viewing bacteria in 3D that have had immunogold labeling applied to their

surfaces, to better localize the location of the gold.

Good luck!

Gib

Gib Ahlstrand, Minnesota Micoscopy Society Newsletter Editor

Electron Optical Facility, University of Minnesota, Dept. Plant Pathology

495 Borlaug Hall, St. Paul, MN. USA. 55108 (612)625-8249

612-625-9728 FAX, giba@puccini.crl.umn.edu

put someone in the audience with stereo glasses, projected the images, and

adjusted until the "audience" saw optimal stereo.

> A special screen was required, but I can't recall the details.

It is one coated with aluminum rather than glass. Metals do not

change the polarization upon reflection, but glass does. I don't know

where they can be obtained; we guard ours jealously.

Yours,

Bill Tivol

tivol@wadsworth.org

not accurate apparent z-values. Depending on what you're going to do with

the info, you may or may not want to use the table. Most computer programs

are capable of working out the correct z-values if you tell them the tilt

angles.

Yours,

Bill Tivol

tivol@wadsworth.org

so,

would you mind describing how the images are lined up?

I 've thought for years that stereo projection had to be one of the most

valuable

assets to any classroom. I remember that the basics included 2 Ektamatic

projectors with crossed sheets of polaroid across their lenses to allow

viewers to

use crossed polaroid glasses. A special screen was required, but I can't

recall the

details. I think I remember that one image was mounted as usual, and the

other

adjusted to the correct stereo location using a special guide that was

invented by

Lee Peachey, I think, and sold by one of the EM suppliers.

That was many years ago, so if there's a better technique now, please let

us hear

about it. And/or correct me if I'm wrong.

Paul D. Millikin, MD

Semi-retired Pathologist

millikin@mtco.com

stereo images and measure heights on them on the web:

http://www.soft-imaging.de/products/modules/m_ste2.htm

and

http://www.soft-imaging.de/products/modules/m_ste.htm

there is also a SEM made salt cystal imaged.

Norbert Overbeck

overbec@uni-muenster.de