ARCHEOGUIDE: First results of an Augmented Reality, Mobile Computing
System in Cultural Heritage Sites
Vassilios Vlahakis††, John Karigiannis†† , Manolis Tsotros††, Michael Gounaris††, Luis Almeida+,
Didier Stricker*, Tim Gleue**, Ioannis T. Christou††, Renzo Carlucci†, Nikos Ioannidis††
*Fraunhofer Institut für Graphische Datenverarbeitung.
**Zentrum für Graphische Datenverarbeitung
+Centro de Computacao Graphica
†A&C 2000 Srl.
This paper presents the ARCHEOGUIDE project
(Augmented Reality-based Cultural Heritage On-site
GUIDE). ARCHEOGUIDE is an IST project, funded by
the EU, aiming at providing a personalized electronic
guide and tour assistant to cultural site visitors. The
system provides on-site help and Augmented Reality
reconstructions of ancient ruins, based on user’s
position and orientation in the cultural site, and real-
time image rendering. It incorporates a multimedia
database of cultural material for on-line access to
cultural data, virtual visits, and restoration information.
It uses multi-modal user interfaces and personalizes the
flow of information to its user’s profile in order to cater
for both professional and recreational users, and for
applications ranging from archaeological research, to
education, multimedia publishing, and cultural tourism.
This paper presents the ARCHEOGUIDE system and
the experiences gained from the evaluation of an initial
prototype by representative user groups at the
archeological site of Olympia, Greece.
Augmented Reality, Mobile Computing, Position
Tracking, Image Rendering, Avatars.
ARCHEOGUIDE is a novel system bringing state-of-
the-art visualization technology and mobile computing
in cultural heritage. It provides personalized Augmented
Reality tours  and reconstructions of ruined cultural
heritage sites in order to help visitors and scientists
better appreciate and enjoy the past glory of these sites.
It features portable units carried by users during their
site tours, communication networks, and a central
database. This multimedia database is the central
depository for the multimedia material making up the
tour presentations. It is also used to support scientific
research and promote cultural education in European
countries and beyond.
The system bridges the gap between recreation and
science and renders culture and history more accessible
to a wider public. Unlike traditional audio commentary
systems widely adopted in cultural sites all over the
world, ARCHEOGUIDE offers a more lively and
realistic experience. The 3D reconstructions of
monuments and artifacts are presented to the user
through a special augmented reality interface while he
has constant visual contact with the natural surroundings
and listens to audio commentary. This feature renders
the system more user-friendly and avoids the
shortcomings of other similar systems where the user is
isolated, or immersed in a purely synthetic world.
Respecting the sensitivity of cultural heritage sites, the
system poses no risk of damage to the site, visual or
visitor disturbance, or interruption of the normal
operation of the site. In addition, it presents a very
simple interface for creating new content and enriching
the database with scientific information.
This paper is organized in 10 sections. The system
architecture is described first and details are given on its
modules. This is followed by a presentation of the user
evaluation of the system and an analysis of these results.
Emphasis is put on presenting a complete description of
ARCHEOGUIDE for the benefit of the reader and other
technical projects in cultural heritage. Details on the
system setup and use scenarios are given and
conclusions are drawn for use in the future system
2. Overview of System’s Architecture
A description of ARCHEOGUIDE’s architecture is
contained in . In this paper we provide more
detailed information for the purpose of completeness
and for helping the reader better understand the user
feedback following the evaluation of the system.
ARCHEOGUIDE is based on a client-server
architecture. It comprises three modules: the Site
Information Server, the Mobile Units, and the
Communication Infrastructure. These modules are
depicted in Figure 1 and analyzed below.
2.1. The Site Information Server
The Site Information Server, or server for short, is a
central PC comprising a multimedia relational database.
It serves as the main depository of the system, storing
audio-visual and textual information covering the site
and its monuments, as well as, related artifacts and
information. The database is organized thematically and
Figure 1 System Architecture.
supports querying according to attributes like
geographic position (coordinates), item type (e.g.
temple), time (e.g. chronological era), level of
information (e.g. general, detail, scientific), theme (e.g.
relates to sports), and versioning (e.g. stage in the
restoration process). Its contents can be accessed by the
mobile units who request material according to user
preferences, position, and selected tour.
A Database Manager (DBM) is in charge of the
operation of the database. It exports a remote access
interface for communication with mobile units and other
servers (e.g. Web servers, museum servers) and allows
querying and retrieval of its contents.
The server is accompanied by authoring tools. These
tools can be used by the site manager for creating new
multimedia content, organizing it, and defining new
tours in the site. The same tools can also be used for
creating and updating existing guidebooks and CD-
ROMS. Alternatively, they provide archeologists and
scientists with a user-friendly graphical interface for
documenting the contents of the site and update it with
the latest research in the field. Figure 2 illustrates the
graphical user interface (GUI) of the authoring tools. It
is based on Windows and supports all standard features
and operations familiar to computer users.
Figure 2 Authoring Tools GUI.
2.2. The Mobile Unit
The clients are a group of Mobile Units (MU). There
are three different implementations of MUs based on
laptop, pen-PC, and palmtop computers, each providing
similar features according to their processing
The top-of-the-range MU, the laptop-based one, is
built on a Toshiba Satellite 800MHz laptop computer,
with 256 MB RAM. A Differential GPS receiver
(Garmin GPS-35 LVS) and a digital compass (Precision
Navigation TCM2) are hooked on the laptop to provide
position and orientation tracking of the touring user. A
PC camera with automatic luminance adjustment for
capturing live video from the user’s viewpoint, and a
special Augmented Reality (AR) Head-Mounted Display
(HMD) (Sony Glasstron PLM-S700) in the form of a
pair of see-through sunglasses for displaying AR worlds
featuring monument reconstructions on top of their
natural surroundings are also attached to the laptop.
The DGPS receiver collects data from a constellation
of 4-12 satellites to calculate the exact location of the
user. The accuracy of the calculations is improved by a
correction signal transmitted from a DGPS beacon
situated at a precisely known position. This accuracy is
typically measured at less than one meter.
Site Information Server
The compass is based on a digital inclinometer sensor
to measure the user’s heading with an accuracy of 0.2°
and allows for magnetic field measurements and
automatic calibration. This approach provides superior
precision and suffers no interference from
electromagnetic fields in the archaeological site.
The position and orientation information is used by the
MU to determine the position of the user in the
archeological site and the monument ruins he is staring
at. Based on this information, the system matches the
corresponding, calibrated images (retrieved from the site
database and being captured from a slightly different
angle) with the live video captured from the camera in
the user’s heading. The matching is performed by a
phase-correlation image registration technique  used
to calculate a warping transform. This transform is then
used to rotate, translate, and scale a 3D VRML model of
the reconstructed ruins the user stares at, so that it can
be rendered on top of the live video image. The image
matching and rendering currently run at 15 frames/sec
resulting in realistic augmented views even when the
user is moving at a fast pace or turning his head. The
resulting AR image is displayed on the HMD. The user
sees a realistic representation of the monument in its
ancient glory and at the same time he has visual contact
with his surroundings, in contrast with other systems
where he is isolated. Audio commentary is coupled to
the visual presentations, giving descriptions and
The hardware components of the MU are integrated in
a compact way and are carried by the touring user. The
user wears the HMD and a bicycle-type helmet upon
which the camera and the compass are mounted. They
are connected to the laptop and the DGPS receiver,
which are carried in a backpack for protection and easy
The pen-computer implementation of the MU is built
on a Fujitsu Stylistic 3400S flat panel PC with 64 MB
RAM. The machine features a pressure-sensitive screen
where the user can write and interact with it with a
special pen, and DGPS and digital compass devices like
the ones described above. This lighter version of the
MU displays pre-rendered 3D reconstructions of ancient
ruins on its screen, accompanied by information and
audio-visual content of related artifacts (e.g. statues
originally found in the visualized temple and currently
stored in a museum). The presented information is
closely related to the position and orientation of the user
as it is tracked by the system. The visual content is
archived in the site database and downloaded to the MU
together with the accompanying audio commentary.
Navigation help is also offered to the user. A digital map
(catopsis) of the site is displayed on screen together with
his current position and the direction he is heading for.
Figure 3 illustrates the map. The heading of the user is
indicated by a red-yellow compass updated in real time.
The lightest version of the MU is built on a Compaq
iPAQ Colour Pocket PC. This is a lightweight and very
compact standalone device supporting user interaction
via a special pen operating on its screen. It presents the
user with similar information as the pen-computer but it
is the user’s responsibility to request the material
corresponding to the monument he is looking at as the
system has no means of determining it automatically.
Figure 3 Navigation Interface.
The user can consult an on-screen digital map to
navigate in the site. However, his position and
orientation are not displayed. This choice (of not using a
DGPS receiver or a digital compass) is based on the
compactness of the device, which would otherwise have
to be sacrificed.
2.3. Communication infrastructure
ARCHEOGUIDE uses two wireless networks for the
communication needs of its components. The first one is
used for the communication between the Mobile Units
and the Site Information Server allowing the transfer of
the audio-visual content of the site tours. It is a wireless
LAN based on the IEEE 802.11 Wireless LAN
Standard. The mobile units and the server can hook up
into the WLAN with a Lucent ORiNOCO wireless
network PCMCIA card and accompanying antenna
plugged into their PCMCIA slots. The network uses a
group of 3 Access Nodes (comprising antenna and
transceiver modules) to fully cover the archaeological
site of Olympia. The traffic handled by the network
includes images, VRML models, and audio clips and up
to 50 mobile units can connect to the network at any
The second network is a point-to-point wireless link
used by a DGPS reference station to communicate
correction signals to the MU’s DGPS receiver. This
station is a LF DGPS beacon transmitting data over the
wireless link in the 300kHz band under the RTCM SC-
3. Overview of System’s Software
This section gives a brief description of the software
components of the SIS and the MU. Server components
have been implemented in Java for maximum
portability, whereas mobile unit components have been
developed in C++ for maximum speed, as they must
operate in real time.
3.1. SIS Software Components
The Site Information Server consists of the following
components: Data Manager, Authoring Tool, and GIS
The Data Manager (DBMS) is built on Oracle 8 and
supports querying and interfacing of the database with
the other system components. The multimedia material
is stored (in several standard formats) outside the
DBMS itself to avoid having to manage a very “heavy”
database. It is stored in a special directory of the
server’s network file system, which requires special
privileges for browsing and/or editing, and is pointed to
by a URL.
The Authoring Tool is used by content creators in
order to insert, delete and manipulate objects in the
multimedia database. Through this interface users can
traverse the archeological site’s database structure in a
tree view, change this structure, manipulate multimedia
objects, group multimedia objects into composite
objects, list multimedia and composites to construct
scripts and, finally, create schedules from existing
The GIS Authoring Tool (GISAT) is a module of the
ARCHEOGUIDE Authoring Tool. This module enables
the user to perform the following operations:
- Visualize the archaeological area in Olympia, as
shown from above, using an orthophotomap of the
site in high resolution.
- Perform basic graphical operations on the image,
such as zoom-in, zoom-out, and distance measures.
- Recognize the coordinates of each point in the
image by a mouse click. Since the image is
registered in both geographical and projection
coordinates systems, each mouse click is identified
by two coordinate pairs: (longitude, latitude) and
(X, Y), in WGS84 and UTM systems, respectively.
- Define new and manipulate existing site node
entities, which reside on the Server Data Manager.
Specifically, the user is provided with Select, Insert,
Delete and Query options to define or manipulate
the geometry of the Server Site Nodes.
- Define new and manipulate existing viewpoint
entities, which reside on the Server Data Manager.
Specifically, the user is provided with Select, Insert,
Delete and Query options to define or manipulate
the geometry of Server Viewpoints.
The SIS tools are accessed via a standard windows
Graphical User Interface, depicted in Figure 2. Standard
operations like Create, Cut, Paste, Save, Edit are
supported with the contents listed in a tree-like structure
for easier visualization. The root of the tree comprises
the whole site with nodes corresponding to smaller
regions (e.g. the Temple of Zeus), and leaf nodes to
multimedia objects (e.g. images).
3.2. Mobile Unit Software Components
The Mobile Unit consists of the following
components: MPS Controller, Image Tracker,
AVALON renderer, and Graphical User Interface
(GUI). All these client components operate
concurrently, with the controller instructing
continuously the renderer what to render based on the
data read from the tracking system, and the user
interactions. The controller’s interaction with the image
tracker is done indirectly via the image renderer. The
controller is also responsible for requesting multimedia
objects from the site database over the wireless network.
So, the MU is implemented as a multithreaded process
with the Multimedia Presentation Synchronization unit
(or Controller for short) being the main thread of
control. Other threads include the handler for the image
tracker, the AVALON renderer, and the GUI.
Figure 4 Tour Schedule Hierarchy.
An important concept is that of the personalization of
information and the hierarchical structure of tour scripts.
As mentioned earlier, the multimedia objects are stored
in the site database together with several attributes.
These attributes are used to select these objects and
create tours adapted to the user’s profile. This profile is
created by the user at the beginning of the tour when he
is prompted by the controller to reply to simple
questions scoring his interests, background, age, sex,
and language of preference. This way, the tour is
adapted to cover regions fitting his description (e.g.
emphasis on sports), the available time for his visit, the
detail and the level of information (implied by his
background and age), and the language of choice. Based
on these parameters, the objects that constitute a tour are
hierarchically ordered as shown in Figure 4. At the
finest level of granularity, there are audio, image, and
VRML objects . These are grouped to form
composite objects, containing the actual data to be
rendered along with additional profile meta-data
characterizing it. For example, a VRML object contains
a string describing the URL location of the VRML file
containing a VRML scene – e.g. the temple of Zeus –
plus data about the Earth co-ordinates of the reference
point of the VRML, plus a direction vector (together
with scale information if necessary). The controller
evaluates the condition associated with each object. If
met, it instructs the appropriate renderer to present it. At
a higher level of complexity, tour scripts represent an
ordered sequence of multimedia objects, all of which are
to be presented if the script is presented. Scripts may
have partial order constraints, which essentially implies
that the content creator may require the visitor to view
one script before presenting another. Scripts are also
logically associated with a geographical area. The
rendering of the objects of each such script is the
responsibility of the renderer and is accomplished with
the help of the position and orientation tracking system.
Finally, at the top of the hierarchy, we find tour
schedules that represent partially ordered sequences of
scripts. This structure aims at helping the controller
synchronize the presentation of these scripts.
The image tracker is the component responsible for
matching the image seen by the user (and captured by
the camera) and the reference images stored in the site
database, and calculating the transformation needed to
render the VRML models on the video sequence ,
, , , , , , , .
The transformation parameters are passed to
AVALON, which performs the rendering at speeds up to
The interaction between the user and the system is
supported by a Graphical User Interface (GUI). Three
different versions exist corresponding to the various
implementations of the MU.
In the laptop-version, the flow of information is
controlled by effectively treating the user as an active
pointing device. Instead of using a mouse, he positions
himself to predefined points-of-interest (or viewpoints)
in the site. His position and orientation triggers the
presentation of the corresponding audio-visual material
on the HMD. The user can intervene and alter the flow
of information by skipping a presentation (by moving
away, or heading to an other direction), or requesting
additional information by means of an optional
In the pen-computer version, the user takes a similar
active pointer role and interacts by means of a special
pen used to “write” on the computer screen. He can
view augmented reconstructions with a standard web
browser, listen to related audio clips, and request
In the palmtop implementation of the MU, the user
can see augmented pictures, read related information,
and listen to audio commentary through a pocket web
browser. No tracking of his location is available and he
is responsible for selecting the material of his interest.
It should be noted that all the multimedia material is
downloaded from the SIS database unless it already
exists in the MU database (e.g. downloaded by the
previous user of the device).
An example of the User Interface presented by the
pen-computer is illustrated in Figure 5. A similar one
exists for the palmtop device, whereas the laptop user
sees only the augmented images on the HMD.
Figure 5 User Interface of the pen-computer MU.
4. Augmented Reality Presentations
The ARCHEOGUIDE system presents its user with
Augmented Reality views of reconstructed ruins. The
calibrated 3D models of the reconstructed monuments
are rendered on the video or still image (according to
the version of the MU) to construct augmented images
like the ones shown in Figure 6.
A special category of these presentations is made
available to users situated inside the stadium. This
category includes the reconstruction of Ancient Olympic
Games with avatar athletes competing in the stadium.
In order to achieve this goal, virtual human models
have been created and animated with high realism and
historical accuracy in terms of the necessary movements
(and/or artifacts) to execute each sporting modality.
Special care was taken in modeling and animating the
virtual athletes (Figure 7), based on specific
bibliography and historical descriptions, published by
international experts on ancient Greek civilization,
about the ancient Olympic athletes and games .
The problem of creating virtual humans in a 3D
environment addresses two issues: modeling the
geometry of the human body and simulating the correct
behaviour and movements. The details of the modeling
process can be considered out of the scope of this paper.
Nevertheless, our approach is based on the
customization of a pre-existing 3D human body, in order
to achieve a model with ancient Greek athlete look
(Figure 7), and model the movements it should describe,
according to the bibliography, for each sport .
Defining human character behaviour includes its
animation to execute “human like” movements .
Methods for animating virtual characters can be roughly
divided into two approaches: motion-based captures and
simulation-based. Motion capture is typically done off-
line and played at runtime. The drawbacks of motion
capture include the relatively high cost of capturing
data, the lack of flexibility of pre-recording, and the fact
that the motion must be performed by someone .
Simulation relies on computation of a mathematical
model to generate the motion of the 3D model. This can
be done on-line or off-line. If the model is too “heavy”,
i.e. the existing hardware is not able to compute the
minimum number of frames per second to achieve
smooth motion, the animation should be calculated off-
(a) Original Image
(b) Augmented Image
Figure 6 Augmented Reality presentation.
Figure 7 Models of ancient Greek athletes.
line, recorded and presented as a video clip. Otherwise,
the animation can be calculated in real time, enabling
some control over the movement parameters. Taking
advantage of such control, it is somehow easy to adapt a
simulation to include variations on the animation
parameters, such as limb, length, or walking speed. This
can be very useful in implementing animations that vary
in time according to the stimuli received from the user
or other objects in the scene. , , and  contain
useful texts on modeling, animation and simulation.
At this stage of development, it is possible to use the
ARCHEOGUIDE system prototype to present the
Diaulos (two-stadium-long race), the javelin throw, the
discus throw, the race in armour and the long jump,
augmenting the corresponding animated models in the
real scenario. To implement the animations of collective
sports, like the Diaulos, the same model was replicated,
differentiating each one by changing its muscular mass
and the time duration of the associated animation. This
way, it is possible to achieve a more realistic
representation for these kinds of sports. To implement
the simulation the H-ANIM  specification was used,
which intends to be a standard representation for human
characters. Figure 8 illustrates a frame of the avatar
video presented to system users at the stadium of
Olympia. The competing avatar athletes are rendered on
the live camera image of the stadium. Realism is
enhanced by the individual behaviour of each athlete
and the fact that the user is not isolated but can, instead,
see other visitors present in his field of view.
An additional feature aimed at giving more realism to
the presentations of the pen-PC and palmtop-based
versions of the MUs is the augmented panorama. This
presentation features 360º views of the natural
surroundings with augmented reconstructions rendered
on top. This enables the user to have a more realistic
experience from the selected viewpoints. In the case of
the pen-PC, the DGPS receiver and the digital compass
readings are used to automatically align and rotate the
panorama with the position and directions the user is
staring at. This, effectively, minimizes the user’s
interaction with the MU and provides real-time
panoramic views. In the palmtop-based MU, the user
can click a direction button to rotate the panorama. This
is necessary, as this unit features no position and
orientation tracking mechanism. These augmented
panoramas feature high quality reconstructions, as is the
case for the standard augmented images. Accuracy and
realism is enhanced by occlusions and shading of the
5. Trials Preparation Procedure
The system was evaluated at the archaeological site of
Olympia, in Greece, by representative user groups. This
section describes the procedure followed for the
First, a site survey was performed to collect
topographical data and was combined with photographic
images used to create a 3-D representation of the site.
Monuments of interest were identified and the temple of
Hera, the temple of Zeus, the Philippeion, and the
Stadium were selected for the initial prototype scenario.
Figure 8 Avatar athletes competing in the Stadium.
Based on this list of monuments and the site
representation, and taking into account occlusions, tour
paths were defined that allowed visitors to view these
monuments. This way, tree-covered areas have been
avoided and optimum GPS data reception has been
achieved by locating an adequate number of satellites
(between 4 and 12). Similarly, candidate locations have
been identified for the installation of the wireless LAN
antennae. Three points have been selected outside the
perimeter of the site, providing good radio coverage to
the selected path and minimum esthetical disturbance.
Second, viewpoints have been selected and marked in
the site. These are locations providing the best views of
the selected monuments. The user is guided by the
system to stand on these points for visualizing the
Third, GPS reception was tested. The DGPS device
provided an accuracy of less than 1 meter.
Fourth, the digital compass was calibrated and tested
to ensure that no strong magnetic field is present that
can compromise the accuracy of its readings.
Fifth, a set of photographs has been taken from the
selected viewpoints and from slightly different angles.
These photographs are used as references in the image-
Sixth, a set of 3-D reconstruction models of temples
has been designed based on archaeological and
architectural reconstruction data. These models
correspond to the reference images and are rendered on
the live video or the still images (depending on
prototype version) to create the augmented views.
A group of 50 representative users was selected. It
comprised visitors of various ages, sex, academic
background, nationality, interests, and computer skills.
This selection ensured that all use scenarios would be
tested and feedback would cover all features of the
system. Emphasis was given in ensuring that the
disturbance to the normal operation of the
archaeological site was kept to a minimum.
The visitors were given a briefing on the system and
assistance during its use. At the end of their tour, they
filled in a questionnaire scoring their impressions on the
system and commenting items they thought significant.
The questionnaire starts with a section recording the
user’s profile and comprises 5 sections. The first three
sections deal with the three implementations of the
mobile units (laptop, pen-PC, palmtop-based), and the
last two sections with general questions and comments,
respectively. A scoring scheme is used where the user
can score between 1 (lowest) and 5 (highest). For
general questions a choice between Yes, No, Not Sure is
The first three sections test for:
- User’s comfort in using each system
- Clarity of audio-visual output (audio descriptions,
augmentations, related objects, animations)
- Realism of audio-visual presentations
- Efficiency and ease-of-use of the multi-modal user
- Flow of information
- Suitability of information
- User’s satisfaction of his tour with each system
- Would the user use the system again?
The fourth section tests the overall user’s impression
of ARCHEOGUIDE and compares the three versions
- User satisfaction
- How informative each version is
Finally, the last section logs user comments in order to
identify in depth his impressions on the
ARCHEOGUIDE touring experience.
The general impression of the users was that
ARCHEOGUIDE is a good system and an interesting
application that they would like to see in other cultural
sites too. They were happy to use it and considered it a
useful learning tool that enhanced their visit. They were,
in general, satisfied by the type of AR, 3D, video and
The most enthusiastic of all were visitors of younger
ages, who can be identified as those with the highest
level of computer skills. These visitors are accustomed
to computer graphics and viewed ARCHEOGUIDE
mainly as a video game or leisure activity.
As we move to older ages with fewer, if any, computer
skills, the system is perceived as a leisure and learning
experience. There is some amount of uncertainty on its
use from computer illiterate subjects, though these
problems were greatly overcome with the help of the
engineers who participated in the tests.
Elderly citizens were interested in testing the system
but the majority of them felt uneasy when first
approaching it and curious to check its capabilities.
People with limited understanding of foreign
languages (mainly elderly and southern Europeans) were
dissatisfied by the limited language support (only Greek
and English) of the prototype.
The most serious complaint came from the use of the
pen-computer-based MU. All users pointed out the low
contrast of its screen, making it very difficult to view it
under direct sunlight.
Further on, the limited available multimedia content
(only the temples of Zeus and Hera, the Philippeion, and
the Stadium) was a negative factor as only part of the
site was covered. Consequently, the system could be
used only in the four most important monuments; other
areas were not supported.
A serious concern deals with the pricing policy for the
rental of the equipment and its insurance in case of
damage or accident.
For the top-of-the-range version, the realistic AR
presentations were highly appreciated but objections
arose on the size and weight of the equipment.
The pen-computer-based MU was rated low due to its
low visibility display but was appreciated for its smaller
dimensions and weight, as well as the fact that it
comprised a single unit.
The palmtop-based MU scored high in its user’s
preferences for its very compact and lightweight design.
The screen was clearly visible under all lighting
conditions, including direct sunlight, and its size was a
reasonable price to pay for easy carrying it in the site.
Both the pen-PC and the palmtop MU received good
comments for the augmented panoramas they feature.
Finally, concern was raised on the availability and the
wait time to rent and use a mobile unit.
The results of the user evaluation of the
ARCHEOGUIDE system at the test site of Olympia
were very encouraging. The system was enthusiastically
accepted by visitors, especially those of younger ages.
Several aspects remain to be solved. However, we
should stress that most of the negative comments were
expected as the prototype under evaluation exhibits a
limited number of features compared to the final system.
Current work focuses on the enrichment of the
multimedia database with high quality, scientifically
accurate material covering all areas and monuments of
the test site, as well as, most major European languages.
Linked to that, is the further development of the
personalization features used in the adaptation of the
tour to the user profile. Additional work is being
performed in improving the hardware of the mobile
units regarding its size and weight, and the visibility of
the portable screens under all lighting conditions. Tests
at other cultural sites are planned.
We would like to acknowledge the EU who selected
ARCHEOGUIDE to be financially supported within the
IST program under contract number IST-1999-11306.
9. Consortium Members
The ARCHEOGUIDE consortium is made up of the
Intracom S.A. (Greece), Fraunhofer Institute of
Computer Graphics (IGD) (Germany), the Computer
Graphics Center (ZGDV) (Germany), the Centro de
Computacao Graphica (CCG) (Portugal), A&C 2000 Srl
(Italy), Post Reality S.A. (Greece), and the Hellenic
Ministry of Culture (Greece).
 R. Azuma, “A Survey of Augmented Reality”,
Proc. SIGGRAPH 95, Course Notes #9
(Developing Advanced Virtual Reality
Applications), ACM Press, 1995.
 N. I. Badler, C. B. Phillips, B. Lynn Webber.
Simulating Humans, Oxford Univ. Press, 1993.
 L.G. Brown. “A Survey of Image Registration
Techniques”, ACM Computing Surveys, 24(4),
pages 325-376, December 1992.
 G. Christopoulos, J. Bastias, The Olympic
Games in Ancient Greece, Ekdotike Athenon,
 H-ANIM Specification for a standard VRML
humanoid, version 1.0, 1998, on-line paper
 R. Holloway, “Registration Errors in
Augmented Reality Systems”, Ph.D.
dissertation, University of North Carolina at
Chapel Hill, 1995.
 W. Press, S. Teukolsky, W. Vetterling, B.
Flannery. Numerical Recipes in C: The Art of
Scientific Computing. Cambridge University
Press, Cambridge, UK, 2nd edition, 1992.
 P. Ratner. 3-D human modelling and
animation,. John Wiley and sons, 1998.
 B. Reddy, B. Chatterji. “An FFT-based
Technique for Translation, Rotation, and Scale-
invariant Image Registration”, IEEE IP, 5(8),
 F. Seibert, “Augmented Reality by Using
Uncalibrated Optical Tracking”, Computers &
Graphics 23 (1999), Special Issue on
Augmented Reality, pp 801-804.
 A. State, G. Hirota, D. T. Chen, W. F. Garrett,
M. A. Livingston, “Superior Augmented
Reality Registration by Integrating Landmark
Tracking and Magnetic Tracking”, Proc. of
SIGGRAPH 96, ACM Press, 1996.
 D. Stricker, P. Daehne, F. Seibert, I. T.
Christou, R. Carlucci, N. Ioannidis, “Design &
Development Issues in ARCHEOGUIDE: An
Augmented Reality based Cultural Heritage
On-site Guide”, Proc. Intl. Conf. on
Augmented & Virtual Environments and 3D
Imaging” (EuroImage 2001), Myconos,
Greece, 30 May-01 June 2001.
 D. Stricker, J. Karigiannis, I. T. Christou, T.
Gleue, N. Ioannidis, “Augmented Reality for
Visitors of Cultural Heritage Sites”, Proc. Intl.
Conf. on Artistic, Cultural and Scientific
Aspects of Experimental Media Spaces (CAST
01), Bonn, Germany, 21-22 September 2001.
 R. Szeliski, H-Y. Shum. “Creating Full View
Panoramic Mosaics and Environment Maps”,
Proc. ACM SIGGRAPH 97, August 1997.
 N. Magnenat Thalmann, D. Thalmann, editors.
Interactive computer animation, Prentice Hall,
 VRML Specification on-line
 M. Wray, V. Belrose; “Avatar in Living
Space”, Proc. VRML 99 of the 4th symposium
on the virtual reality modelling language, 1999.
 I. Zoghiami, O. Faugeras, R. Deriche. “Using
Geometric Corners to Build a 2D Mosaic from
a Set of Images”, CVPR97, 1997.