J Neurosurg 95:1057–1066, 2001

Anaglyphic three-dimensional stereoscopic printing:

revival of an old method for anatomical and surgical
teaching and reporting
Technical note

GUILHERME CARVALHAL RIBAS, M.D., RICARDO FERREIRA BENTO, M.D.,

AND ALDO JUNQUEIRA RODRIGUES, JR., M.D., PH.D.

Clinical Anatomy, Department of Surgery, University of São Paulo Medical School, São Paulo,
Brazil; Department of Neurological Surgery, University of Virginia, Charlottesville, Virginia; and
Hospital Israelita Albert Einstein, São Paulo, Brazil

! The authors describe how to use the three-dimensional (3D) anaglyphic method to produce stereoscopic prints for
anatomical and surgical teaching and reports preparation by using currently available nonprofessional photographic
and computer methods.
As with any other method of producing stereoscopic images, the anaglyphic procedure is based on the superimposi-
tion of two slightly different images of the object to be reproduced, one seen more from a left-sided point of view and
the other seen more from a right-sided point of view. The pictures are obtained using a single camera, which follow-
ing the first shot can be slid along a special bar for the second shot, or by using two cameras affixed to a surgical micro-
scope. After the images have been distinguished from each other by applying different complementary color dyes, the
images are scanned and superimposed on each other with the aid of nonprofessional imaging-manipulation software
used on a standard personal computer (PC), and are printed using a standard printer. To be seen stereoscopically, glass-
es with colored lenses, normally one red and one blue, have to be used.
Stereoscopic 3D anaglyphic prints can be produced using standard photographic and PC equipment; after some
training, the prints can be easily reproduced without significant cost and are particularly helpful to disclose the 3D char-
acter of anatomical structures.

KEY WORDS • microanatomy • microsurgery • stereoscopy •

surgical anatomy • three-dimensional imaging

I
N spite of living in a 3D world, our teaching and report-
ing methods are fundamentally based on the written
word and on 2D images. When first-hand experience
is unavailable, understanding of a previously unseen and
complex structure or environment is usually achieved on-
ly by reference to a previous comprehension of its major
planes and not by presentation of an original 3D reproduc-
tion, in part due to the difficulty encountered in depicting
it properly. On the other hand, familiarity with any given
structure or environment is definitely related to a profound
understanding of its 3D character.3,13
Surgical and radiological identifications of structures are
particularly based on the knowledge of anatomy as seen
from and through natural spaces, and hence their shapes
and spatial distribution assume a major significance for sur-
gical and radiological orientation. In the nervous system
the complexity of cerebrospinal fluid spaces and neural
Abbreviations used in this paper: ASA = American Standards
Association; f-stop = maximum flash-to-subject distance for flash
photography; f(followed by numeral) = size of lens aperture open-
ing; PC = personal computer; SLR = single lens reflex; 2D = two-
dimensional; 3D = three-dimensional.
structures increase our difficulty in understanding the 3D
character of the system.
To disclose the tridimensionality of anatomical spaces
and structures, we began to produce stereo slides of fixed
brain dissections and surgeries for stereo projection, which
already had been done by others in the past.1,2 The success
we have experienced using this teaching tool during the last
3 years recently led us also to produce 3D stereoscopic
prints for the same aim. The established anaglyphic meth-
od, which was initially described by Almeida, as cited in
Ferwerda,3,18 was chosen by us because it is the simplest
one for printing purposes.
Material and Methods
Basic Overview
All available methods of producing stereoscopic images
are based on the initial procurement of at least (and usual-
ly) two slightly different images of the object to be repro-
duced. These two images differ because one demonstrates
the object as seen more from a left-sided viewpoint and
the other shows the object as seen more from a right-sided

J. Neurosurg. / Volume 95 / December, 2001 1057


viewpoint, in a manner that corresponds to how each of our
eyes would individually view that object.
These two images are known as a stereopair and can be
obtained using two cameras placed slightly apart or by slid-
ing a single camera across a slide bar while photographing
still objects. The distance between the two cameras must be
proportional to normal human interpupillary distance and
the distance from the spectator to the object location. The
simplest rule is to calculate the distance between cameras
as being 1/30 of the distance between each individual cam-
era and the object to be photographed. The apparent change
in position of an object resulting from a change in the posi-
tion of the viewer or camera is a phenomenon called “paral-
lax.”3,18
The second image of the stereopair can also be obtained
using an appropriate computer-editing process, in which
a standard 2D scanned picture is transformed with the aid
of software specially designed for imaging manipulation,
such as Adobe Photoshop. The development of such im-
ages requires skill and familiarity, both with the use of the
software and with stereoscopy itself, to create the appropri-
ate parallaxes.
To view the reproduced object stereoscopically, each im-
age of the stereopair has to be seen only by one eye of the
observer, which is achieved with the aid of special viewers
or by utilization of different polarizing modalities and spe-
cial glasses. Glasses composed of simple polarizing filters
are called “passive” glasses; glasses equipped with liquid
crystal technology, which are capable of briefly alternating
occlusion of one eye while the other eye sees its respective
image of a given stereopair, are called “active” or “shutter”
glasses. Stereoscopic video is based on the same principles
of stereo photography, and any projection of superimposed
polarized images requires a projection screen covered by
metal paint, usually aluminum, which keeps the reflected
light polarized and is known as a silver screen.3,18
For printing purposes the anaglyphic method presented
in this article constitutes the method of choice because of
its simplicity and low cost.
The Anaglyphic Method
The anaglyphic method was initially developed by Al-
meida in 1858 for stereo projection and is based on the the-
ory of distinguishing each image of the stereopair by apply-
ing a different complementary color dye, usually red and
blue or red and green, with red generally being assigned
to the left eye. The procedure can also be used to obtain
3D stereoscopic prints composed of superimposed red- and
blue-dyed images of a given stereopair, which can be seen
with the aid of red-and-blue filter glasses.3,18
The rationale of the anaglyphic method is the fact that
each complementary color filter removes the components
of an image that were previously dyed with that color,
which almost leaves for the eye’s vision only its comple-
mentary component, recreating a situation in which binoc-
ular vision can focus on a given object from two distinct
points of view.
Recent computing developments have made the ana-
glyphic method much easier. Nowadays, the original ste-
reopair images can easily be scanned digitally, superim-
posed with the aid of imaging software, and printed with
the aid of standard color printers.
1058
G. C. Ribas, R. F. Bento, and A. J. Rodrigues, Jr.
The final anaglyph can be created in a gray-scale mode,
which appears as a black and white picture when the view-
er dons the red-and-blue filter glasses, or can be fashioned
in a color mode, although there is significant sacrifice of
the complementary colors used as filters. When creating a
colored anaglyph, the colors have to be carefully chosen,
with red being the most likely to be sacrificed. When using
red and blue dyes, the blue still maintains some of its tonal-
ity, and green, yellow, brown, and white are stronger colors
as a result. Using dark blue and yellow as separating colors
is better if one wishes to preserve all colors; however, this
variation is very uncomfortable for the spectator because
one eye has to view the object through a very dark filter.
The anaglyphic method can also be used with comput-
er monitors, slide projections, and video presentations. Its
projection does not require a silver screen, but has the same
color limitations already mentioned, which makes it signif-
icantly inferior to dual projections of polarized full-color
slide stereopairs.
Technical Details
Photographs of our anatomical dissections were obtained
using a Nikon FM2 SLR camera and a 105 mm 1:4 Mi-
cro-Nikkor lens adapted to a regular ring flash. The film se-
lected was standard 100 ASA Kodak Ektachrome daylight
film. Most photographs were obtained from a distance of
approximately 90 cm at a speed of 1/60 seconds and using
f8 or f11. The camera was affixed to a Jasper sliding bar,
which was adapted to a standard tripod. The camera’s hori-
zontal displacement was calculated according to the afore-
mentioned 1/30 rule.
Photomicrographs of anatomical dissections and surgical
views were obtained using two Nikon N2000 cameras di-
rectly affixed to both side extensions of a Carl Zeiss surgi-
cal microscope. The anatomical dissections were photo-
graphed through two f220 extensions adapted to the right
and left sides of a Carl Zeiss OPMI MDM surgical micro-
scope, and the intraoperative pictures were obtained with
the aid of two f85/340 dual adapters connected to a Carl
Zeiss OPMI CS-NC Contraves surgical microscope.
Both cameras should be very carefully aligned so that
they will not be rotated in relation to each other.
Although the halogen light source of the first microscope
required 160 ASA Tungsten Ektachrome film, the light
source of the OPMI CS-NC Contraves surgical microscope
required 100 ASA Ektachrome daylight film. All photomi-
crographs were obtained using a 1-second exposure, with
f-stops depending on the amount of light and zoom enlarge-
ment. Trial and error involving different exposures is al-
ways necessary to achieve better photographs.
After the photographs had been obtained, they were
digitally scanned using a Nikon LS 2000 scanner, and the
anaglyphic images were created with the aid of Adobe
Photoshop or VRex Depth Charge software by using a 200-
MHz Pentium-processor computer with 128 Mb of random
access memory and a 4-Gb hard disk connected to a Sony
monitor. Printing on standard glossy paper was performed
using an Epson color printer.
The lenses contained in the anaglyphic glasses should be
the proper color to ensure cancellation of unwanted imag-
es; they can be obtained from special 3D merchants.
J. Neurosurg. / Volume 95 / December, 2001
Anaglyphic three-dimensional stereoscopic printing
FIG. 1. Anterior transcallosal 3D macroscopic view of the anterior horns of the lateral ventricles.
Sources of Supplies and Equipment
The FM2 SLR camera and the two N2000 cameras, as
well as the 105-mm 1:4 Micro-Nikkor lens, and LS 2000
scanner were obtained from Nikon Corp. (Tokyo, Japan).
The 100 ASA Ektachrome daylight film and the 160 ASA
Tungsten Ektachrome film were purchased from Eastman
Kodak Co. (Rochester, NY). The sliding bar was obtained
from Jasper Engineering (Mountain View, CA). The OPMI
MDM and OPMI CS-NC Contraves surgical microscopes,
the f220 extensions, and the f85/340 dual adaptors were ac-
quired from Carl Zeiss (Oberkochen, Germany).
Adobe Photoshop (version 5.0 for Windows) was manu-
factured by Adobe Systems, Inc. (San Jose, CA) and VRex
Depth Charge software by VRex, Inc. (Elmsford, NY). The
J. Neurosurg. / Volume 95 / December, 2001
210 SF Multiscan Triniton monitor was manufactured by
Sony Corp. (Tokyo, Japan). The 1520 Stylus Epson color
printer was obtained from Seiko Epson Corp. (Nagano-
Ken, Japan). Anaglyphic glasses can be obtained from Reel
3-D Enterprises, Inc. (Culver City, CA).
Illustrative Images
A selection of stereoscopic pictures is displayed in
this article to demonstrate the various possibilities of the
method.
Figures 1, 2, and 3 are 3D macroscopic images of the
brain obtained by sliding a single camera along a slide bar.
These three images disclose the main neural relationships
of the anterior horn and the body of the lateral ventricles.
1059
 G. C. Ribas, R. F. Bento, and A. J. Rodrigues, Jr.

FIG. 2. Three-dimensional macroscopic view of the relationships between the lateral ventricle and the cerebral medi-
al surface, temporal operculum, insula, and subinsular structures.

Figure 4 demonstrates the possibility of obtaining a 3D
stereoscopic image from a greater distance by using two
cameras attached to each other. This photograph demon-
strates the microsurgical setting in the operating room.
Figure 5 is a surgical stereoscopic image of a superi-
or cerebellar artery aneurysm obtained with two cameras,
each one attached to one side of the surgical microscope.
The cameras should be carefully positioned so that the im-
ages viewed are identically aligned.
Figure 6 demonstrates how a surgical procedure can be
stereoscopically visualized step by step, and how this meth-
od can be used properly to disclose the 3D character of
anatomical spaces and structures. These pictures exhibit
the main anatomical relationships encountered during an
anterior temporal lobectomy and amygdalohippocampec-
tomy.17
Discussion
Biological evolution led to humans being a fundamental-
ly visually oriented species, mostly to the detriment of our
olfactory function. The binocular vision provided by the
proper alignment of eyes and orbital cavities is one of the
major characteristics of primates and makes possible the
observation of a single object from two slightly different
points of view, which constitutes the basic feature of the 3D
or stereoscopic visual perception that is processed in our
brains.
Despite the fact that the disparity of images seen by right
and left eyes constitutes the essential factor of true ste-
reoscopic vision, a significant depth perception can also be
obtained by single-eye vision through so-called monocular
cues, such as movement of objects, their relative size, lin-
ear perspective, texture gradients, shadows, and occluded
aspects, among other previously learned visual characteris-
tics.13,18 These monocular cues, of course, also enhance bi-
nocular stereoscopic vision.
Refined creative and motor activities not only require ac-
curate 3D vision, but also the development of mental abil-
ities that enable us to be constantly spatially oriented and

1060 J. Neurosurg. / Volume 95 / December, 2001

Anaglyphic three-dimensional stereoscopic printing
FIG. 3. Three-dimensional macroscopic view of the main relationships between the anterior horn and anterior com-
missure.
able to build spatially based strategies. In the case of sur-
gery, these sorts of mental activities are not only based on
the simple knowledge the surgeon has of surgical anatomy
but are particularly related to the ability the individual pos-
sesses in dealing with 3D mental images.
The possibility of reproducing surgical anatomy pictures
J. Neurosurg. / Volume 95 / December, 2001
and sketches with 3D stereoscopic images for teaching
and reporting purposes clearly enhances the actual illus-
tration by properly displaying the depths of forms and
shapes. Contrary to so-called 3D computer-generated imag-
es, which are displayed on 2D surfaces and in which the
perception of depth is only enhanced through utilization of
1061
 G. C. Ribas, R. F. Bento, and A. J. Rodrigues, Jr.
FIG. 4. Three-dimensional image demonstrating the microsurgical setting. The image was obtained using two cameras
affixed to the surgical microscope.
perspectival methods and other monocular cues already
mentioned, 3D stereoscopic images provide a real depth
perception that mimicks direct human vision.
Three-dimensional stereoscopic anatomical images are
already available, mostly for visualization through proper
viewers1,2 and computer monitors.15,16 Authors of the classic
Bassett Stereoscopic Atlas, originally published in 19611
and recently reissued,2 took advantage of Sawyer View
Master reels and viewers and a 2D printed guidebook.
These images can be projected though a special View Mas-
ter projector onto a silver screen, and the original images
can also be transformed into 35 mm slides for polarized
dual projection. Two microsurgical stereo atlases, one edit-
ed by Poletti and Ojemann in 198510 and the other edited by
Kraus and Bailey in 19945 contain fine 3D stereoscopic im-
ages, which are also displayed on View Master reels that
can be seen using the proper viewer.
Recently, anatomical and surgical stereoscopic video and
still images have been used for teaching purposes mostly
through projections and through computer monitors. The
1062
printing of stereoscopic images has not been used regular-
ly for medical purposes, however, although it is feasible by
using procedures already known for many decades. Other
than the uncomfortable crossing-eye or parallel-eye meth-
od required to look at a stereopair to see a third converged
3D image, the simplest way of viewing such images is
through the aforementioned anaglyphic method.
From the 1890s up to the 1950s, there were many print-
ed applications of two-color anaglyphs in magazines and
papers, particularly in 3D comics publications in the early
1950s. Recently, in the August 1998 issue of National Geo-
graphic Magazine there appeared two reports in which the
red-and-blue anaglyphic method was used to display 3D
stereoscopic images of explorations of Mars8 and the sunk-
en ship Titanic.6 In São Paulo, Brazil, a major newspaper
has also begun to publish some 3D stereoscopic advertise-
ments by using this method under our guidance.9
The lenticular procedure3,18 constitutes an alternative
method for stereoscopic printing, but it is definitely more
difficult, more expensive, and less adequate to show details
J. Neurosurg. / Volume 95 / December, 2001
Anaglyphic three-dimensional stereoscopic printing
FIG. 5. Three-dimensional microsurgical view of a superior cerebellar artery (SCA) aneurysm.
than the anaglyphic method. Lenticular photographs should
be obtained by combining at least four superimposed pic-
tures reflecting different points of view and covered by a
surface composed of very narrow cylindrical plastic lenses
(lenticules), usually narrower than 0.3 mm. Due to the cy-
lindrical shape of its surface, the interlaced images can be
viewed stereoscopically without any optical aid. Lenticular
postcards were particularly popular in the 1960s, and dur-
ing the 1980s and early 1990s Nimslo and Nishika lenticu-
lar four-lens cameras were commercially available.14
Besides providing a better anatomical illustration and
improved 3D comprehension, the utilization of stereoscop-
ic images might also enhance the individual’s familiar-
ity with tridimensionality itself and favor spatially related
abilities. Gestalt psychologists4,13 emphasize that form is
the primitive unit of perception and stereoscopic images
hold the definite advantage in producing their illustration.
Because surgery is actually performed through natural
spaces and through the development and enlargement
of virtual spaces that exist between structures, the ideal
knowledge of surgical anatomy clearly requires 3D under-
standing of these anatomical relationships. This intuitive
anatomical framework is basic for a better neuroimaging
interpretation and for proper microsurgical and endoscopic
surgery orientation.
Virtual environments for planning, training, and per-
J. Neurosurg. / Volume 95 / December, 2001
forming surgery, which are being developed through tele-
surgical systems,7,11,12 will also require stereoscopic visual-
ization and the individual’s familiarity with 3D conceptual-
ization and 3D stereoscopy.
Last, but not least, medical illustration and documenta-
tion is a field in which science meets art. The development
of well-constructed 3D stereoscopic images may also con-
stitute another way of expressing the beauty of anatomy.
Conclusions
Besides providing attractive and elegant images, and
producing an unusual and fun experience both to create and
to see, 3D anaglyphic prints constitute a particularly realis-
tic teaching and reporting tool especially for surgeons and
radiologists, because they emphasize the spatial distribu-
tion that characterizes real anatomy. Nowadays, the proce-
dure can be performed using standard photographic and PC
equipment and can easily be reproduced.
Acknowledgments
We are grateful to the Department of Otorhinolaryngology at the
University of São Paulo Medical School and to the Otorhinolaryn-
gology Foundation of São Paulo, for their support, and to Mr. Ron
Labbe for his technical assistance.
1063
 G. C. Ribas, R. F. Bento, and A. J. Rodrigues, Jr.

FIG. 6. Three-dimensional image demonstrating the anatomical relationships encountered during an anterior temporal
lobectomy and amygdalohippocampectomy. A: View of the hippocampus, fimbria, and choroid plexus in the temporal
horn after removal of the temporal lobe neocortex and opening of the choroidal fissure, with exposure of the posterior
cerebral artery in the ambient cistern. B: View of the brain after anterior dissection was performed just in front of the
head of the hippocampus and medial dissection was performed along the opening of the choroidal fissure. The choroid
plexus is located to the left of the thalamus and the lateral geniculate body is just above the posterior cerebral artery.
FIG. 6. (continued)➝

1064 J. Neurosurg. / Volume 95 / December, 2001

Anaglyphic three-dimensional stereoscopic printing

C: Exposure of the cerebral peduncle, the posterior cerebral artery in the ambient cistern, and the remaining neural
structures after posterior dissection and hippocampectomy. The basal vein of Rosenthal is hidden behind the choroid
plexus in the roof of the ambient cistern. D: View of the removed segment. On the medial side the hippocampal sul-
cus lies between the subiculum (part of the parahippocampal gyrus) and the dentate gyrus (part of the hippocampus),
which in this view is covered by fimbria. FIG. 6. (continued)➝

J. Neurosurg. / Volume 95 / December, 2001 1065

 G. C. Ribas, R. F. Bento, and A. J. Rodrigues, Jr.

E: Final view of neural structures after hippocampectomy and removal of both the uncus and the portion of the amygda-
la contained within the uncus. This view was achieved by sectioning the anterior aspect of the temporal stem. Superi-
orly, the amygdala is continuous with the basal aspect of the globus pallidus, with both structures together assuming a
dumbbell shape.

References
1. Bassett DL: A Stereoscopic Atlas of Human Anatomy. Port-
land, OR: Sawyer, 1961
2. Chase RA: A Stereoscopic Atlas of Human Anatomy, The
Bassett and Gruber Legacy. Borger, The Netherlands: 3D
Book Productions, 1994
3. Ferwerda JG: The World of 3D, ed 2. Borger, The Nether-
lands: 3D Book Productions, 1990
4. Kofka K: Principles of Gestalt Psychology. New York: Har-
court Brace Jovanovich, 1935
5. Kraus GE, Bailey GJ: Microsurgical Anatomy of the Brain:
A Stereo Atlas. Baltimore: Williams & Wilkins, 1994
6. MacInnis JB, Kristof E: Titanic: tragedy in three dimensions.
National Geographic 194:120–127, 1998
7. Mahoney DP: Brain storm. Computer Graphics World 21:
79–88, July, 1998
8. Newcott WR: Return to Mars. National Geographic 194:2–29,
1998
9. O Estado De São Paulo Lança Anúncio Em 3D. O Estado de
São Paulo:A5, A7, A9, A15, A17, A19, B9, D7, and E4, May
11, 1999
10. Poletti CE, Ojemann RG: Stereo Atlas of Operative Micro-
neurosurgery. St. Louis: Mosby, 1985
11. Satava RM: Virtual reality, telesurgery, and the new world or-
der of medicine. J Image Guided Surg 1:12–16, 1995
12. Satava RM: The virtual surgeon. The Sciences 38:34–39, 1998
13. Solso RL: Cognition and Visual Arts. Cambridge: MIT Press,
1994
14. Starkman D, Pinsky S: The Nimslo 3D Book. Duarte, CA: Reel
3D Enterprises, 1986
15. Trelease RB: Toward virtual anatomy: a stereoscopic 3-D in-
teractive multimedia computer program for cranial osteology.
Clin Anat 9:269–272, 1996
16. Trelease RB: The virtual anatomy practical: a stereoscopic 3D
interactive multimedia computer examination program. Clin
Anat 11:89–94, 1998
17. Wen HT, Rhoton AL Jr, de Oliveira E, et al: Microsurgical
anatomy of the temporal lobe: Part 1: mesial temporal lobe
anatomy and its vascular relationships as applied to amygdalo-
hippocampectomy. Neurosurgery 45:549–592, 1999
18. Wolfgram DE: Adventuras em 3D. São Paulo: Berkeley Brasil
Editora, 1993
Manuscript received March 21, 2000.
Accepted in final form July 25, 2000.
Address reprint requests to: Guilherme Carvalhal Ribas, M.D.,
R. Eduardo Monteiro 567, Sao Paulo 05614–120, Brazil.

1066 J. Neurosurg. / Volume 95 / December, 2001