SCCH 20092nd Conference Scientific Computing and Cultural Heritage

BOOK OF ABSTRACTS

November 16th – 18th, 2009Heidelberg University, BIOQUANT, INF 267

SCCH 2009

Organizing Committee

Hans Georg Bock Willi Jäger Hubert Mara Michael J. Winckler

Cover design

Elisabeth Pangels

komplus GmbH

Copy Editing

Elisabeth Trinkl

Heidelberg Graduate School of Mathematical and Computational Methods for the Sciences

Prof. Dr. Dr. h.c. Hans Georg BockInterdisciplinary Center for Scientific ComputingHeidelberg University Im Neuenheimer Feld 36869120 Heidelberg, Germany

Dr. Michael J. WincklerInterdisciplinary Center for Scientific ComputingHeidelberg University Im Neuenheimer Feld 36869120 Heidelberg, GermanyT + 49 - 6221 54 - 4981email Michael.Winckler@iwr.uni-heidelberg.de

Oktavia Klassen / Ria Hillenbrand-LynottIm Neuenheimer Feld 368, Room 50769120 Heidelberg, GermanyT + 49 - 6221 54 - 4944email hgs@iwr.uni-heidelberg.de

Office Hours for HGS MathComp: Mon – Thu 09:00 – 12:00 & 13:00 – 16:00

Helpdesk for applications etc.: Tue, Thu 10:00 – 12:00

Chairman

Administrative Director

Office

BOOK OF ABSTRACTS

SCCH 2009

2nd Conference

Scientific Computing and

Cultural Heritage

November 16th – 18th, 2009

Heidelberg University

SCHEDULE

Monday,

November 16th

Tuesday,

November 17th

Wednesday,

November 18th

09:00-

10:40

Invited Lecture

Session 3

09:00-

10:40

Invited Lecture

Session 5

10:30-

11:15Registration

10:40-

11:00Coffee Break

10:40-

11:00Coffee Break

11:15-

12:30

Opening

Invited Lecture

11:00-

12:20Session 4

12:30-

14:00Lunch Break

12:20-

14:00Lunch Break

11:00-

12:20

Session 6

Best Student

Presentation Award

Closing Remarks

14:00-

15:40Session 1

15:30-

16:10Coffee Break

16:10-

17:30Session 2

17:30-

19:00

Poster Session

Welcome Reception

14:00-

18:30

Guided Tour:

City of Speyer

Address

BIOQUANT

Im Neuenheimer Feld 267

Room SR 041 and SR 042

Ruprecht-Karls-Universität

Heidelberg

69120 Heidelberg

Contact

Oktavia Klassen

Email: scch2009@iwr.uni-

heidelberg.de

Phone: ++49-6221-54-4944

4

How to get to BIOQUANT

By car

• Coming from the motorway

Turn left towards "Chirurgie", cross the Neckar on Ernst-Waltz-

Brücke and follow Berliner Straße till the 2nd

traffic light (in front of

the Shell petrol station). There turn left to the institutes.

• Coming from Neckargemünd

Follow "Uferstraße", then turn into "Posseltstraße" ("Jahnstraße"

respectively); at the traffic light turn right into "Berliner Straße" and

follow the road till the 2nd

traffic light (in front of the "Shell" petrol

station). There turn left to the institutes.

By public transport

• From city center (Bismarckplatz)

Take bus number 31 towards Neuenheim, "Chirurgische Klinik" and

get out at "Technologiepark".

You can also take tram number 21 towards Weinheim and get out at

"Bunsengymnasium".

• From main station

Take tram number 21 towards Handschuhsheim, "OEG-Bahnhof" or

tram number 24 towards Handschuhsheim, "Burgstraße" and get out

at "Bunsengymnasium".

Social events

Welcome Reception – Monday, November 16th

Time: 5.30 pm

Location: BIOQUANT

Im Neuenheimer Feld 267

Ground Floor

Guided Tour: City of Speyer – Tuesday, November 17th

Departure in Heidelberg

Time: 2.00 pm

Location: in front of the BIOQUANT building

Departure in Speyer

Time: 6.00 pm

ABSTRACTS

Session 1

6

BAATZ W., FORNASIER M., HASKOVEC J., SCHÖNLIEB C.-B.

Mathematical methods for spectral image reconstruction

In old frescos, the visible colour information might be completely or

partially lost in some parts of the painting. This is due to specific

chemical reactions of the pigments, which modify their absorption of

visible light. However, if these reactions do not largely influence the

absorption of the pigments in invisible parts of the spectra (UV and

IR), there is a hope that the original colour information can be

faithfully recovered, using the information from the well conserved

parts of the painting. We demonstrate how mathematical methods

for sparse matrix recovery can be used for this task. As shown by

Candès and Recht (2009), the missing data can be exactly recon-

structed with very high probability (i. e., for “almost all” matrices),

given only a mild lower bound on the number of sampled entries.

Quite recently, two numerical algorithms have been proposed for

sparse matrix recovery: The singular value thresholding (SVT)

algorithm by Cai, Candès and Shen (2008), and the iteratively re-

weighted least squares minimization (IRWLSM) by Daubechies,

DeVore, Fornasier and Güntürk (2009). In addition to these two

algorithms, which are iterative in nature, we propose a third method

(block completion, BC) for recovery of the missing elements of a low-

rank matrix, which, although based on a trivial algebraic

manipulation, delivers very competitive results. However, this

method can only be used if the matrix rows and columns can be

permuted such that the missing elements constitute a block; in our

case this is always possible.

To demonstrate the performance of these three methods, we use a

sample painting consisting of linear combinations or red, yellow and

blue (Fig. 1, top). We divide the painting into a grid of 20×20

rectangles, and on each of these rectangles we measure the

absorption spectra in the range 307 – 1130 nm; a typical human eye

will respond to wavelengths from about 380 to 750 nm, wavelengths

below 380 nm and, resp., above 750 nm correspond to UV, resp. IR.

From the visible spectral data, we reconstruct a rough approximation

of the original painting (Fig. 1, bottom). For our experiment, we pick

randomly a certain portion of the rectangles (50%) and delete the

visible parts of the measured spectra, while keeping the UV and IR

Session 1

7

regions (Fig. 2, top). We represent these data as a matrix, where

each row corresponds to one measured spectra, and test the

performance of the three algorithms (SVT, IRWLSM, BC) for recovery

of the deleted information. The performance is measured in terms of

the relative error of the recovered visible spectra with respect to the

original data. We show that the SVT algorithm typically reaches a

relative error of approx. 30% before the convergence drastically

slows down, while the IRWLSM usually goes down to 20% or even

better. The BC method typically does even better (10%), and,

moreover, has the advantage of being extremely simple to

implement and non-iterative, and, therefore, very fast. The result

after applying the BC method is shown in Fig. 2, bottom. Finally, we

make the experiment of deleting the whole lower part of the

painting, such that the information about the visible spectra

corresponding to the blues is completely lost (Fig. 3, top).

Surprisingly, it was possible to recover the missing parts quite well

(Fig. 3, bottom), where again the IRWLSM and BC methods gave

best results.

Fig. 1: Sample fresco

painting (top) and its

projection on the 20 × 20

rectangular grid (bottom)

Fig. 2: Random deletion of

50% of the elements (top)

and recovery by the BC

method (bottom)

Fig. 3: Deletion of the

lower part of the image

(top) and recovery by the

BC method (bottom)

Poster

8

BALZANI M., FEO R. De, VANUCCI C., VIROLI F., ROSSATO L.

The Angel’s cave. A database for the restoration and

valorisation of the San Michele Archangel site, Olevano sul

Tusciano (Salerno, Italy)

Along the Picentini mountains slopes the Mt. Raione houses the

entrance of San Michele Archangel cave. The place was used since

the Neolithic period but the first historical data are linked to the IX

siècle when it became a natural shelter for the bishop Pietro and,

later, venue of pilgrimage. Due to the presence into the cave and its

branches of bizantinian frescos, a church and some Martiryas (little

chapels) with small courtyard the sanctuary is an unique example of

important religious cave in Italy. Recently, an archaeological

campaign found out interesting ceramics objects such as the

medieval ceramic Forum Ware made by roman traditions.

Extraordinary ancient music instruments, the Tibiae, were also found

into the cave: they were made carving shinbones and then used as

flutes by local inhabitants in ritual ceremonies. After a joint effort of

the Soprintendenza per i Beni Architettonici e per il Paesaggio, il

Patrimonio Storico, Artistico e Demoetnoantropologico per le

Province di Salerno e Avellino and the centre DIAPReM

(Development of Integrated Automatic Procedures for Restoration of

Monuments) of the Department of Architecture of the University of

Ferrara, laser scanning integrated technologies were used in order to

obtain a first survey aiming to show the huge quality of the site

trough a complete documentation action. The research project was

finalized to give a strong base for the restoration and valorisation of

the San Michele Site and the surrounding landscape; in the

meanwhile it was a good opportunity to verify an integrated survey

process in a low accessibility area in order to evaluate:

1 – the feasibility level of a such extreme condition

technological survey;

2 – the instrumental acquisition degree of definition in

relations to the morphometric level of detail;

3 – how the survey could help to the configuration of a

comparative model aiming to show the degradation process

and the loss or modification of the extraordinary architectural

and artistic heritage;

Poster

9

4 – how the morphometric database could be enquired in

order to define further scenarios of conservation and

valorisation of the site.

The survey started with the main branch of the cave where a laser

scanner Leica HDS 3000 (based on a time-flight technology which

allow data collection of big volumetric complexes acquiring circa

1000 points per second with an accuracy of 6 mm) was used to

obtain the needed information. The three-dimensional data were

then integrated by a topographic survey to realize a model made of

55.000.000 acquired points by the which was possible to drawn up

the cave plans, sections and façades and a scaled plaster model. The

output will be useful to build a structured collection of records

organized on several layers thought for information exchange,

divulgation and for the realisation of revitalization project of this

extraordinary site.

Poster

10

BALZANI M., MAIETTI F., GALVANI G., SANTOPUOLI N.

The 3D morphometric survey as efficient tool for documen-

tation and restoration in Pompeii: the research project of Via

dell’Abbondanza

The project titled From Asellina to Verecundus: research, restoration,

and monitoring addressing painting on certain famous Pompeian

botteghe in Via dell’Abbondanza (Regio IX, Insulae 7 and 11) was

characterised by close collaboration between the Soprintendenza

Archeologica di Pompei, the “Valle Giulia” School of Architecture at

the University of Rome “La Sapienza”, the School of Architecture and

the DIAPReM Centre of the University of Ferrara, and the School of

Engineering II of the University of Bologna (Forlì campus). Its

primary objectives were the safeguarding of famous architectural re-

mains and experimentation with restoration methodologies and

materials.

The restoration work addressed a number of façades along the

stretch of the Decumanus Maxiumus between the Forum and the

Sarnus Gate (a stretch known today as via dell’Abbondanza). The

façades were unearthed in 1912 during excavation work under the

direction of Spinazzola.

After the collection of numerous notes from previous archaeological

investigations and from visual inspections about architectural

morphology, materials and state of conservation, surveys of ancient

façades were carried out and measurement data were collected. The

survey by means of 3D laser scanner of the varied and complex

architectures have been characterized by an attempt to focus efforts

on contributing representational knowledge of the existing site

elements.

The comparison between Pompeii as it once was and Pompeii as it

now is and observations of how it changes, decays, and mutates day

by day offer an extraordinary arena for experimentation and re-

search. The nearly ten years of research and experimentation at

Pompeii have been characterised by an attempt to focus efforts on

contributing representational knowledge of existing site elements.

The chosen research field, which nevertheless remains open to inter-

disciplinary approaches, is archaeological excavation and restoration

and problems of conserving our cultural heritage.

Poster

11

Cross section of Via dell’Abbondanza and long section of the House of Paquius Pro-

culus obtained by the 3D database. The survey context involved via dell’Abbondanza

and the house of Paquius Proculus, on the opposite side of the restoration project

area. The survey of the entire exteriors and interiors has been performed mainly by

means of laser scanners using time-of-flight technology. These scanners are able to

rapidly acquire a high definition 3D point cloud (accurate to 5–6 mm) and are highly

reliable instruments on the monumental and urban scale.

The newly developed technologies for the automatic acquisition of

geometrical data are innovative elements that allow us to create

databases of high definition, three-dimensional morphometric data.

These archives of architectural and archaeological data are a

valuable research tool for archaeologists, architects, and historians of

art and architecture, but also, and above all, they serve the purpose

of protecting and conserving cultural heritage sites and provide sup-

port to restoration processes and training programmes. The data-

base contains 3D models obtained by use of the laser scanner and

all the topographical, photographic, diagnostic, and structural data

associated with them. The database allows users to consult and

update all data. This provides an important foundation for the

management, conservation, maintenance, and enhancement of

Pompeii’s extensive, complex, and diversified urban, architectural,

and monumental legacy.

The experimentation addressed critical historical aspects, restoration

methods and materials, and the protection and maintenance of

painted and architectural surfaces. It’s our opinion that the critically

study of Pompeii still shows in a manner that is in some ways unique

the history of the methods of archaeology and restoration, up to the

use of the most modern technologies that in certain case are truly

transforming it in a kind of advanced laboratory.

Session 3

12

BOOS S., MÜLLER H., HORNUNG S.

A multimedia museum application based upon a landscape

embedded digital 3D model of an ancient settlement

In its traditional sense, cultural heritage, whether tangible or

intangible, can be defined as monuments, cultural and natural sites,

museum collections, archives, manuscripts, etc., or practices that a

society inherits from its past, and which it intends to preserve and

transmit to future generations. Digital technologies in this regard

increasingly assume a high significance due to their contribution in

digital preservation and the abilities for three dimensional digital

reconstructions of cultural assets. In terms of a sustainable digital

preservation the development of common principles and standards

for the handling of digital content play an important role. Initiatives

like the London Charter, which aims on establishing internationally

recognised principles for the use of three-dimensional visualisation

by researchers, educators and cultural heritage organisations, or the

development of the OGC 3D modelling standard City GML exactly

pursue these goals.

Considering the principles of the London Charter this abstract

describes the development of a digital reconstruction of the celtic

hillfort “Altburg” (Germany), which was generated in the context of a

museums exhibition in the Hunsrück-Museum Simmern (Germany)

and which refers to the City GML standard.

The spoken to exhibition highlights the art of living of the regional

iron-age cultural group “Hunsrueck-Eifel-Culture” (HEK), which

denotes iron-age tribes of the Hunsrueck and Eifel mountains in the

West of Germany. Especially during the 5th and 4th century BC

these regions became increasingly important in far-ranging trade or

political connections, archaeologically detectable in precious

imported grave-furnishings. As a result of these processes of social

development a considerable number of hillforts was constructed to

protect people or supplies in times of crisis. Almost all archaeological

evidence of the HEK derives from graves and has to be judged on

the background of varying burial rituals. In contrast hardly anything

is known about settlement activities from that time, because hardly

anything else than post-holes from wooden houses or simple pits

have survived the centuries. Since no visible traces above ground

Session 3

13

remain from these iron-age farmsteads, it is very hard to locate

them and therefore only few sites are known. For that reason it

became clear that in order to convey a picture of the circumstances

of life about 2500 years ago, the exhibition relies mainly on

reconstructions.

Besides reconstructions of contemporary objects like costumes,

jewellery, an authentic replica of a post-built granary and cinematic

re-enactments of life and craftsmanship in Celtic times, a multi-

media based application serves as a platform for detailed information

about the HEK. In this regard a 3D-model of the celtic hillfort

“Altburg” near Bundenbach (RhinelandPalatinate) was developed, a

site belonging to the best preserved remains of that kind. Several

excavation campaigns in the 1970ies could ascertain four building

phases of the hillfort and due to the excellent preservation of the

site, which is manifested in the remains of postholes and ditches in

the subsoil a very detailed image of the hillfort could be derived. In

the first building phase (ca. 300 BC) the Altburg consisted of few

larger houses, where its inhabitants lived, a number of granaries for

food-storage and a round-house of maybe public character, whose

precise purpose is not yet known. The settlement was surrounded by

a simple wooden palisade guarded by a fortified gateway. Since

archaeological sources are outstanding, the community of

Bundenbach decided to reconstruct the settlement of one of the later

settlement phases at the original location, using even the excavated

post-holes for the buildings.

Unfortunately this reconstruction is neither complete, nor does it

succeed in conveying an authentic impression of the iron-age

settlement. The valley below is nowadays densely wooded, so that

the Altburg seems remote, but back in the iron-age all trees and

undergrowth would have been cleared to make the settlement a

visible landmark. Therefore the only way to convey an impression of

the earliest settlement seems to create a virtual 3D-model of its

building phase I.

With regard to the London Charter and the requirements of the City

GML standard, which defines several Levels of Detail (LOD) for multi-

scale 3D modelling, the decision was taken to define both the

landscape model and the model containing the ancient buildings as

close as possible according to the CityGML standard. The imple-

mentation, however, was done by using different off-the-shelf

Session 3

14

software. Thus in a first step the digital landscape and the digital

reference for the 3D models of the buildings is generated. The input

data consist of:

• Digital Elevation Model (DEM) with a resolution of 10 m

• Ortho image with a ground resolution of 20 cm

• True scaled finding plan in a 1/400 scale

The ortho image is used for georeferencing the finding plan and the

plan in turn is used to assess the positions of the single buildings by

creating point features, which are positioned in the center of the

digital footprints of the buildings. The results of these steps are

visualised in the 2,5 environment of the software ESRI ArcScene.

Finally the hillfort buildings, which are constructed and textured in

the 3D sketching software Google Sketchup are imported as 3D

marker symbols and are adjusted to their orientation and the

topographical situation (Fig.).

Fig.: 3D Model of the celtic hillfort “Altburg”

In order to acquaint the museums visitors closer with the historical

scenery an animation was generated in ArcScene in the form of a

retrospective virtual flight over the landscape and around the mount,

where the hillfort was situated. To achieve this objective the land

use presentation sequentially changes from today’s situation into the

presumed ancient celtic time. Afterwards transitions as well as

Session 3

15

textual information in the form of a lead text, subtitles and end titles

are added by use of the software Windows Movie Maker. Additionally

a painted representation of the hillfort scenery was created and

appended to the animation in form of a slow cross fade after the last

frame of the animation to the artistic representation. Finally the

video product is integrated into an overall multimedia presentation

about the HEK, which is developed by means of HTML and

JavaScript techniques.

Session 3

16

BROSCHART J.

Data, Science, and the Media

There is a growing interest for science among people outside of

academia. This has positive and negative side effects. While it is

good for scientists to relate to the general public in order to account

for their work and the money they get for it, there is often a clash

between scientific interests and what appeals to laypersons. As a

consequence, what eventually gets published in the media is

frequently unsatisfactory from a scientist's point of view. On the

other hand, it is often the case that the scientists in question fail to

provide the relevant information. In his talk, Jürgen Broschart, who

works as editor of the science news section of GEO magazine, will

illustrate the basic pitfalls in science-media interaction. With respect

to the topic of the conference, good and bad examples of data

presentation will be discussed, and what can be done to enhance the

aesthetic appeal of scientific data.

Session 5

17

CHRIST G.

A Collaborative GIS-based Cue Card System for the

Humanities

The junior research group “Trading Diasporas in the Mediterranean

1350–1450” and the other groups in the Special Research Program

“Transcultural Studies” at the University of Heidelberg is involved in

the development of the software litlink, a multi-role database system

for the management of literature, different types of sources, meta

information and research design. This paper explains the importance

of such a system sketching out the history of knowledge manage-

ment in the humanities since the early modern times. Then, it gives

an account of what has been achieved so far and an outline of what

is planned for the future: The development of a server-based col-

laborative multi-role research environment with an advanced GIS

interface.

Problem

Human society changes over place and time. This is the realm of

historical research. Historians map historical change over place and

time. Traditionally, they did this on paper and on large format maps

familiar from geography and history classes.

Historical research is based on two key ingredients: sources (com-

prising: documents, pictures, objects, archaeological evidence) and

research literature (books, articles etc.). Traditionally, historians

stored this material either in a system of notebooks or on cue cards

(cf. the notebook system (loci method) of Erasmus of Rotterdam or

the cue card systems of Niklas Luhmann or Umberto Eco). The cue

card systems of the latter reach a high level of sophistication:

Besides cards with bibliographical data, excerpts and citations, the

system comprised thematical cards, author cards and workflow

cards. With this system of cards link to each other by a sophisticated

keyword-system expressed in a code of numbers and letters, he

succeeded in creating a very efficient and highly intuitive system of

knowledge management. His system gained its own dynamism

growing up to a certain size and could produce links and innuendos

the individual researcher’s brain would not be able to produce.

Session 5

18

However, since then, the introduction of IT in the humanities

generally seemed to have led to a regress rather than a progress in

knowledge management. Word processers replaced the cue card

system and information was now stored again in serial (MS-word)

files. Furthermore, lack of geographical analysis and reluctance to

work collaboratively hamper historical research furthermore.

Solution

We tested different IT solutions to remedy this unsatisfactory

situation: We found that on the one hand there is bibliographical

software: most popular ranks the commercial software endnotes. On

the other hand, there are several smaller projects, which focus on

the classical cue card functions: storing and linking ideas intuitively.

Only one software tries to combine the two functions, bibliographical

database (with the possibility to retrieve information from library

catalogues etc.) and cue cards. This is the filemaker-based freeware

litlink (www.litlink.ch). Litlink is a cooperative project between Prof.

P. Sarasin, PD Dr. Haber and Nicolaus Busch (programmer) and,

increasingly, our research group at the University of Heidelberg.

So far, the system is able to process

• Literature

• Archival documents

• Objects

• Events

• Persons/Groups

Session 5

19

• Cue cards linked to all the categories above

• Ideas and Projects

Thus, litlink in its present form allows not only for the storing of

bibliographical information, archival sources, objects and the

respective excerpts and notes but also of places, persons

(prosopographical database), ideas and projects, linked in an

intuitive system based on a routine comparing the different items by

their keywords.

Perspectives-GIS

There is also server-version of litlink. As it stands, it is the same

database system as in the stand-alone version: There is neither the

possibility to define differentiated access rights, say for guests,

standard users and administrators nor to differentiate the rights of

use: reading only, reading and writing, etc. This is a problem: For

the humanity’s research culture tends to be highly conservative and

suspicious, the individual researcher is reluctant to upload his

material on a shared database unless he can control it. He wants to

define exactly who can see, comment or even edit “his” items.

Consequently, we are planning currently how to overcome these

obstacles and to gear up litlink adequately for use by a research

group and its connected knowledge community.

Since the interaction with other databases as Jstor, library

catalogues etc. or with other bibliographical software as endnotes is

presently rather weak, we aim at improving litlink’s functionality

honing its interactive capabilities with such databases.

Currently, litlink is able to store places with the respective

geographical information in order to visualize a given output.

Extracting a set of data, say the list of events forming the biography

of a historical personality, in a GIS system (for instance) google

earth the user can plot the data on a map.

The aim is to improve the geographical representation by importing

it into more sophisticated state-of-the-art GIS software as for

instance GRASS. Maybe, eventually, it might prove helpful to rebuild

the whole database system (currently realized in litlink) on such a

GIS base. This re-engineered would at the same token transfer the

research environment into a state-of-the-art sql-database-system.

This would allow for the integration of web 2.0 features and further

improvement of collaborative research.

Poster

20

FELICIATI P.

MAG, an Italian XML application profile for submission and

transfer of metadata and digitised cultural content

MAG (Metadati amministrativi gestionali, see http://www.iccu.sbn.it/

genera.jsp?id=267) is an Italian metadata application profile totally

compliant to international standards, developed with the main goal

of promoting among cultural organizations the aggregation of a

“least common” set of technical and management metadata to

guarantee the good submission and transfer of metadata and

cultural digital objects (text, images, audio, video) in local or

distributed digital libraries (SIP and DIP phases of OAIS model).

MAG, developed in the framework of the national digitalization plan

“Biblioteca Digitale Italiana” in 2001 by the Central Institute for

Unique Catalogue and is maintained since that time by an ad hoc

Committee. The application profile, presently available in its 2.01

release, was adopted in the last 8 years by many cultural heritage

digitalization projects, especially in Libraries and in Archives.

MAG enables the full use of metadata maintained and defined in

other schemas (Dublin Core and NISO) in association with specific

metadata defined for the particular context of cultural digitalization

projects (where it wasn't possible to find a full answer in existing

profiles). The XML schema covers general informations on the

project and the type of digitalization adopted, descriptive metadata

about the analogical object digitized, structural metadata describing

the logic structure of the digitized object, five sections dedicated to

record technical informations about digital objects (images, audio,

video, text, ocr) and one section developed to collect some

information on objects availability and access in order to enable their

dissemination.

The wide dissemination of MAG standard in Italy created the

conditions for the developing of several applications by software

private companies – conceived to manage cultural heritage digiti-

zation projects in every phase, from data capture to web publication.

As regards the relationships between MAG and other metadata

application profiles, in some cases the Italian standard was used –

correctly as a sort of extension of METS “packaging schema”, a

powerful metadata management tool but with no direct solutions to

Poster

21

some of the technical and practical requirements MAG aims to face

effectively. Moreover, two groups inside the MAG Committee are

presently developing the mapping references MAGMETS and

MAGPREMIS, to guarantee the longterm preservation of MAGbased

cultural digital repositories. The MAG reference document is going to

be published in english in order to be available in September, before

the conference.

The paper will present synthetically the MAG standard's goals, its

structure and elements, presenting indeed some good practices in its

application.

Session 5

22

FERSCHIN P., KULITZ I.

Archaeological Information Systems

Large amounts of data arise out of excavations. The documentation

of this multifaceted information can materialize in different kinds of

media, e.g. photographs and excavation diaries etc. In the course of

the subsequent research work the data will be interpreted, analysed

and supplemented with additional information. At the end of ar-

chaeological research a book is usually published. The texts and

images, such as plans, drawings of findings, diagrams, photographs

etc., of which the book is comprised, mirror the current state of

research as accurately as possible.

If one wishes to have a general idea of the excavation data or create

a digital reconstruction of a building or urban structure, a quite time-

consuming process occurs in which the relevant data from the

publications is extracted and accordingly compiled.

Often experts of diverse disciplines, e.g. archaeology, architecture,

computer science etc., collaborate together on the conception of a

virtual reconstruction using numerous modern digital documentation

methods such as photogrammetry, photography, videos and 3D

documentation techniques. Thus, their information is also incor-

porated into the data and media pool.

Platforms are becoming necessary that make it possible to integrate

the diverse data as well as utilise its visual structure in addition to

simplifying the information and data exchange between distant

researchers.

In this work visual archaeological information systems are being

developed with particular consideration of space and time with the

following goals:

• Concentration of information (“Visual Index“)

• Integration of diverse media

• Publication medium

The purpose of the visualisations is to take the three-dimensional

data from the excavations and restore it to its three-dimensionality

as well as to virtually reconstruct its spatiotemporal basis (four-

dimensionality).

Furthermore, the virtual space should be utilised for the depiction of

additional information. Thus, all the information from the remains,

Session 5

23

such as spatial and temporal references, dimensions and levels,

building materials, finds, excavation photographs, bibliographies etc.,

can be clearly and visually consolidated.

Often the archaeological indices are not present or do not suffice for

a digital reconstruction. Many questions remain open and can be

answered only partially. Therefore, in addition to the depiction of the

findings the following information will be integrated into the

visualisations:

• Differentiation between remains and reconstruction

• Photographs and panoramic images documenting the current

status

• Digital videos with annotations as excavation diaries

• Links to comparative objects

• Digital publications und references

• 2D and 3D-sections

• Variations of reconstruction

• Visual differentiation between certain and uncertain

reconstruction

Examples of archaeological information systems were realised with

Google Earth and 3D-PDF. Google Earth ensures the spatiotemporal

display of information and different media as well as providing a

working and publication platform that could be accessed via the

internet. 3D-PDF allows the easy switching between structured

layers of 3D models to show temporal developments or the archaeo-

logical interpretation and translation of inscriptions. Furthermore it is

very well suited to be included into digital publications.

The archaeological data originated from the excavations of the

German Archaeological Institute / Department of Cairo with which

the IEMAR cooperate in the area of virtual archaeology.

Session 6

24

FLÜGEL Ch., SCHALLER K., UHLIR Ch.

“Archäologische Museen vernetzt” – An Information System

for the Archaeological Museums in Bavaria

The project is based on a common initiative of the Archäologische

Staatssammlung and the Landesstelle für die nichtstaatlichen

Museen in Bayern. It is aimed at the development of an innovative

module-based internet application. Technical concepts, development

and implementation are conducted by CHC – Research Group for

Archaeometry and Cultural Heritage Computing, University of

Salzburg, funding is provided by Bayerische Sparkassenstiftung. The

database driven information system supports the user in exploring

the services made available by the archaeological museums. Based

on a highly flexible data model the unique emphases of the

individual museums will be clearly represented.

Modules of the information system “Archäologische Museen vernetzt”

Testbed Mainlimes

The information system offers spatially structured and harmonized

views on all information offers provided by supra-regional and

regional museums and collections, visualising also the strong ties

between the museums and the extramural (architectural) monu-

ments that are already part of the Limes world heritage.

In a first phase the area covered will be confined to the Mainlimes

territory in Hessen and Northern Bavaria from Groß-Krotzenburg to

Miltenberg. This sector of the Obergermanischer Limes forms a

Session 6

25

testbed to create best practice for designing and operating the inter-

active information system. Later on a diversification to others than

archaeological museums and other regions in Bavaria is planned.

Further on the system will form a link between local museum

internet pages and the information platform offered by

www.museenin-bayern.de, run by the Landesstelle für die nicht-

staatlichen Museen in Bayern.

The archaeological museums at the Mainlimes are an ideal environ-

ment for developing innovative web-based access strategies, as only

a small number of them are strictly confined to archaelogical topics

alone and the information provided by them varies significantly in

terms of quantity and quality.

Basic elements of the information system

• presentation of spatial correlations between museums and

extramural monuments based on highly flexible interactive

cartography

• presentation of all available information material, including

printed matter, (virtual) reconstructions, multimedia based

information and educational services provided by the

museums

• user friendly access provided by theme clustering,

storytelling and stepwise graded depth of information

• integration of external web resources (e.g. the Ubi Erat Lupa

monument database)

• representation of the main focuses of the individual

collections combined with easy to use tools for the creation

of interactive theme clustering and virtual exhibitions

The technical implementation is based on guidelines edited by the

European Union for the presentation of Cultural Heritage content

providing barrier-free access, technical robustness based on open

source technology and avoidance of PlugIns to a possibly wide

extent.

Basic functions like the creation of virtual exhibitions, updating the

calendar of events and integration of multimedia-based content will

be provided by tools that need no expert knowledge.

To provide long term availability and data storage the application will

finally be hosted by the Bavarian State Library in Munich within the

framework of the BLO (Bayerische Landesbibliothek Online) portal.

Session 4

26

GROSMAN L., GOLDMAN T., SMIKT O., SHARON G., SMILANSKY U.

Computerized 3D modeling – a new approach to quantify

post-depositional damage in Paleolithic stone artifacts

The morphological typology of lithic artifacts (handaxes, cleavers,

etc.) is the main tool for following the cognitive development and the

technical skills of the early humans and to distinguish between sites,

regions and cultural phases. Yet various agencies can modify the

artifacts during the long time which elapsed between their deposition

and present day. They can be subjected to e. g., rolling and batter-

ing while in floods and sea waves, or by the action of modern con-

struction and agricultural machinery. Thus, it is imperative to under-

stand quantitatively the damage patterns so that the archaeological

analysis could be (at least statistically) corrected for these effects.

We simulated damaging by battering by placing 8 recently produced

handaxes in a barrel together with basalt pebbles. The damage

occurred by turning the barrel, and it was monitored by scanning the

handaxes in 3D after 5, 10, 20, 40, 60, 100, and 200 turns. Thus,

the complete damage history was recorded. We shall present our

results which quantify the effect of battering on various mor-

phological parameters, and discuss the consequences of our findings

within the methodological context of pre-historical research.

Figure: Damage history of 8 handaxes rolled in the barrel as expressed by their profile

variations

Session 4

27

Our results show that the damage patterns resulting from rolling the

tools in a battering environment are distinct, definable and, most

importantly, different from the patterns of controlled intentional

knapping by humans. Bifacial tools are produced by a rational de-

liberate reduction sequence that involves mainly the use of bifacial

retouch for the homogeneous shaping of the tool edges. Yet after

quantifying the various morphological parameters of the battered

experimental handaxes our findings show that the edges are heavily

battered with sporadic deep concave scars along them and most of

the damage occurs in the area of the handaxe’s tip.

For the archaeologist who specialize in the study of handaxes the

barrel experiment is a “shocking” experience. The resulting breakage

patterns, scar removals, edge retouch and the flakes removed are all

known from many archaeological assemblages studied in the past. In

many ways by following the history of damage many past obser-

vations regarding biface morphology may become questionable.

Session 3

28

HAUCK O., NOBACK A.

Computing the “Holy Wisdom”

The church of Saint Sophia (Holy Wisdom) in Istanbul – formerly

known as Constantinople – was the cathedral of the city. This unique

building with its wide cupola was built by emperor Justinian I (527–

565) between 532 and 537.

The first design by Anthemios of Tralles and Isidor of Milet had to be

changed during the construction phase because of statical problems.

During the following centuries, many windows had to be filled with

brickwork because of structural collapses after several earthquakes.

For the project “The Saint Sophia of Justinian in Constantinople as a

scene of profane and secular performance in late antiquity” which

was funded by the Deutsche Forschungsgemeinschaft (German

Research Foundation) in the framework of the priority programme

“theatricality” a CAD model of this first design of Saint Sophia in

Istanbul has been generated at Technische Universität Darmstadt’s

faculty of architecture. This model is based on the architectural

survey of the American Robert van Nice as well as on personal

inspection of the actual building and the ancient calculating

geometry of Hero Alexandrinus.

Session 3

29

We found out very quickly that for recreating the light effects of the

architectural concept it is as essential to reconstruct the number and

location of the windows as to gather all the surfaces accurately. The

whole building is a highly complex interaction between the occurring

daylight and the window openings, the materials, and even some of

the detail geometry. The vaults which are mainly covered by gold

mosaics are a major component of the light effects. The vaults

reflect the daylight which occurs mostly through the openings of the

aisles into the nave. This was the reason why these vaults but also

all the other surfaces of the internal architecture of the building had

to be reconstructed concerning their original geometry as well as

light reflecting qualities.

We would like to present three major aspects of our work at your

conference:

• The role of ancient mathematics (e.g. Hero Alexandrinus‘

Metrica and Stereometrica) for making the building

computable

• The collection of accurate data and the application of

material descriptions to the computer model

• The use of a lighting simulation software (Radiance) for

imaging.

Session 2

30

HEIN A., MÜLLER N. S., KILIKOGLOU V.

Heating efficiency of archaeological cooking pots: Computer

models and simulation of heat transfer

Studies of archaeological cooking pots revealed a large variety of

vessel shapes and, furthermore, diverse methods of clay paste

preparation, which have been used for their production. Shape-

related parameters concern for example wall thickness, diameter,

curvature and height, while material-related parameters include the

base clay – or base clays in the case of mixing – non-plastic

materials used for tempering and the firing conditions. All these

parameters are known or suspected to affect the performance of

ceramic vessels. In order to be able to address issues, such as raw

material choice and vessel suitability, it is necessary to investigate

how these parameters influence the performance of a cooking

vessel, and which of the performance characteristics are affected

when varying selected parameters.

In former technological studies of archaeological cooking pots,

different performance properties were investigated, mainly focusing

on strength, toughness, thermal shock resistance and the so-called

‘heating effectiveness’. The latter was a first approach to quantify

the heat transfer in cooking pots: replicas of archaeological cooking

pots were manufactured, which were subject to heating experiments

under controlled conditions, examining the rate with which the

temperature of the pot’s content was raised. The ‘heating effective-

ness’ determined in this way, however, is a rather complex para-

meter, depending on thermal conductivity, heat capacity, per-

meability and vessel shape.

This paper presents a novel approach to systematically quantify the

heating efficiency of archaeological ceramics, using computer

models. Digital models of cooking pots were investigated with

numerical methods, such as the finite element method (fig.), which

was recently already applied to other ceramic types. Material data

were used, that had been collected in a recent study on the material

properties of Bronze Age cooking ware. A model of a cooking pot

and its content is exposed to simulated heat loads and the tem-

perature development in the entire system is calculated. The method

allows for straightforward estimation of the heating efficiency, which

Session 2

31

is the ratio of the heating energy which reaches the content and the

heating energy which is applied to the cooking pot. The advantage

of the simulation approach is that models of any cooking pot can be

tested, in terms of ceramic fabric or shape. Furthermore, the

constraints of the cooking process can be freely selected and varied,

such as the temperature of the heat source, temperature of the

environment and properties of the content. Importantly, the

influence of selected parameters on heating rates and times can be

singled out, overcoming some of the shortcomings of the replication

approach described above. Also, since heating efficiency is directly

related to energy efficiency, different makes of cooking vessels can

be compared in terms of their energy consumption.

The present approach allows for the estimation of heating efficiency

and the evaluation of the particular parameters and variation

therein, such as shape and material properties, in a straightforward

way. Furthermore, different kind of cooking processes can be

simulated, allowing the assessment of the suitability of various

cooking pots for different methods of food preparation.

Session 6

32

HOHMANN G., SCHIEMANN B.

Das Projekt WissKI

Wissenschaftliche Forschungsprojekte im Bereich der Museen er-

zeugen ausgehend von den verwalteten Objekten umfangreiche

Sammlungen primärer Grundinformationen, die die Grundlage

jeglicher weitergehender Forschung darstellen. Die Identifikation,

Erschließung und Katalogisierung der Objekte und das inhaltliche

Verlinken der Information bilden die Grundlage musealwissenschaft-

licher Informationsdatenbanken. Das von der Deutschen Forschungs-

gemeinschaft von 2009 bis 2011 geförderte Projekt WissKI

(Wissenschaftliche KommunikationsInfrastruktur) erweitert das Wiki-

Konzept um Komponenten, die es Kuratoren/-innen und

wissenschaftlichen Mitarbeitern/-innen von Museen ermöglichen

sollen, diese Informationen zu erstellen und deren Publikation zu

unterstützen. Die im wissenschaftlichen Arbeitsprozess entstehenden

Stoffsammlungen, die nach außen als Kataloge oder Korpora

(Primärinformation) in Erscheinung treten, sind in aller Regel so

umfangreich, dass aus Kostengründen meist nur eine Auswahl aus

den Daten für den Druck bzw. die Publikation berücksichtigt werden

kann. Ein beträchtlicher Teil der Primärinformation geht mit Ab-

schluss des jeweiligen Projekts für weitere Forschung verloren, da

weitergehende Nutzungskonzepte fehlen. Das im Projekt entwickelte

System soll diesem Problem begegnen, indem es einen

demokratischen Redaktions- und Publikationsprozess für wissen-

schaftliche Information unterstützt, eine Kommunikationsplattform

für die beteiligten Akteure schafft und die Langzeitverfügbarkeit der

gesammelten Informationen ermöglicht. Dies wird u. a. durch die

Berücksichtigung folgender Komponenten gewährleistet: Granulares

Rechte- und Moderationsmanagement, Sicherstellung der Identität

der Autorenschaft, Sicherstellung der Authentizität der Information,

Herstellung der Zitierfähigkeit der Beiträge, Tiefenerschließung durch

Textanalyse und semantischer Repräsentation, Aufbau eines geringe

Kosten verursachenden Veröffentlichungsprozesses. Durch den

dauerhaften Urhebernachweis der Beiträge werden wissenschaftliche

Gratifikationssysteme wie der Citation-Index unterstützt.

Damit eine semantische Verknüpfung dieser Inhalte werkzeug-

unterstützt möglich wird, ist eine inhaltliche Annotation notwendig.

Session 6

33

Zur Realisierung setzt das Projekt daher vollständig auf semantische

Technologien zur Wissensrepräsentation und Tiefenerschließung. Als

Referenzontologie für das gesamte System wird das CIDOC

Conceptual Reference Model (CRM) herangezogen. Das CRM ist eine

Ontologie für den Bereich des Kulturellen Erbes, die seit über 10

Jahren von einer Arbeitsgruppe des International Committee for

Documentation (CIDOC) des International Council of Museums

(ICOM) weiterentwickelt wird. Die aktuelle Version 5.0.1 verzeichnet

86 Entitäten (Klassen) sowie 137 Eigenschaften (Relationen). Seit

das CRM 2006 in der Version 3.4.9 als ISO 21127 zertifiziert wurde,

wird es in kulturwissenschaftlichen Informationssystemen

zunehmend berücksichtigt. Zur Verwendung in Softwaresystemen

liegt das CRM in einer OWL-DL-Implementierung vor, die als

Erlangen CRM (ECRM) in Erlangen von den Projektpartnern ent-

wickelt wurde.

CRM und ECRM verfolgen einen ereignisbasierten Beschreibungs-

ansatz, der es erlaubt, auch die zeitliche Dimension von musealen

Ausstellungsobjekten detailliert zu erfassen. So wird beispielweise

ein Objekt nicht einfach mit einem Herstellungsdatum versehen,

sondern mit einem Herstellungsereignis verknüpft, welches

Ausdehnung in Raum und Zeit haben kann. So lassen sich alle

Ereignisse in der Historie eines Objekts (Modifikation, Verkauf,

Zerstörung etc.) auf einem Zeitstrahl anordnen und mit Personen

und Orten verknüpfen. Im Projekt wird das ECRM als Referenz-

ontologie genutzt, die in einer WissKI-Basisontologie erweitert und

spezifiziert wird. Als Ausgangspunkt zur Modellierung dieser

Basisontologie diente Museumdat10, ein XML-basierter Metadaten-

standard, der zukünftig den Datentransfer zwischen den Museen in

Deutschland vereinheitlichen soll. Museumdat, in weiten Teilen eine

Übernahme des bekannteren Usamerikanischen Standards CDWA-

Lite11, wird im System auch als Import/Export-Format für die Daten

verwendet, die über eine OAI-PMH-Schnittstelle12 kommuniziert

werden. Dieser Import/Export basiert in WissKI auf den seman-

tischen Definitionen der Basisontologie, welche bereits in diesem

frühen Projektstadium als vorläufige Version vorliegt. Weitere

Konzepte und Rollen erweitern die Basisontologie projektspezifisch,

um eine möglichst gute und detaillierte semantische Beschreibung zu

erreichen.

Session 6

34

Eine zentrale Komponente der WissKI-Plattform ist die Verknüpfung

der wissenschaftlichen Texte mit den Konzepten und Rollen der

Ontologien. Dazu ist ein semi-automatischer Annotationsprozess der

natürlichsprachlichen Texte vorgesehen, welcher von den Kura-

toren/-innen und wissenschaftlichen Mitarbeitern/-innen durch-

geführt werden soll. Der WissKI-Prototyp umfasst einen WYSWYG-

Editor für die Eingabe, der zusätzlich als Schnittstelle zwischen den

Analyseverfahren für den natürlichsprachlichen Text und den Be-

nutzern dient. Weitere notwendige oder zusätzliche Annotationen

werden in tabellarischer Form ausgefüllt und stehen damit der Ver-

knüpfung mit anderen Inhalten zur Verfügung. Zur Unterstützung

der automatischen Annotationskomponente und zur Normierung

manueller Eingaben werden etablierte Normdaten (Getty Vokabulare,

Normdaten der Deutschen Nationalbibliothek etc.) teilweise in das

System integriert.

Um die oben aufgezählten Ziele zu erreichen und die verschiedenen

semantischen Integrationsprozesse umzusetzen, bedarf es einer

flexiblen, skalierbaren, frei verfügbaren Basissoftware. Dafür wurde

das Content Management System Drupal ausgewählt. Drupal bietet

neben einer Benutzerschnittstelle, die an Wiki-Software angelehnt

ist, umfangreiche Funktionen (Nutzerverwaltung, Medieneinbindung,

Revisionsmanagement etc.), die über eine einheitliche Schnittstelle

zugreifbar und programmierbar sind. Die projektspezifischen Kom-

ponenten werden als Drupal-Module der Basissoftware beigefügt.

Dazu gehört unter anderem ein einheitlicher RDF-Triple-Store, in

dem sowohl die Ontologien als auch die Annotationsergebnisse

(Instanzen) verwaltet werden, eine umfangreiche OWL-DL/RDF-Ver-

arbeitungschicht inklusive Import/Export und ein WYSWYG-Editor

mit Komponenten zur halbautomatischen Annotation von Freitexten.

Um die Übertragbarkeit der WissKI-Projektergebnisse auf andere

Projekte zu gewährleisten, sind drei unterschiedliche Anwendungs-

szenarien vorgesehen: Der WissKI-Prototyp wird mit den vollstän-

digen Primärdaten eines DFG-Projekts zur Nürnberger Goldschmiede-

kunst, Primärdaten des gerade angelaufenen Projekts zur Dürer-

forschung am GNM sowie mit Primärdaten des BIOTA E15-Projekts

zur Biodiversitätsforschung aus dem ZFMK versehen. Anhand dieser

Szenarien sind verschiedene Tests mit wissenschaftlichen Mit-

arbeitern/-innen geplant.

Poster

35

HOPPE Ch., KRÖMKER S.

Towards an Automated Texturing of Adaptive Meshes from

3D Laser Scanning

Mid range 3D laser scanners become more and more popular espe-

cially for surveying and measuring construction sites, in the field of ar-

chitecture and for preservation of monuments. The data we used

were recorded by a CALLIDUS precision systems CP 3200 using an ex-

tremely short laser pulse emitted by the scanner. The measured time

of flight is proportional to the distance between scanner and object.

Figure 1. Point cloud of the Heidelberg Tun (German: Grosses Fass) in true color RGB-

values.

During the measuring procedure, the laser scanner integrated in the

measuring head rotates by 360º along the horizontal plane with a

resolution of 0.25º. By means of a rotating mirror, the laser beam is

emitted in the shape of vertical fans and thus covers of 140º in the

vertical plane with the same resolution. The measuring shadow on

the ground depends on the height of the measuring head. We deal

with data, which were combined from up to ten or twenty sets of

360º full scans in order to avoid shadows in complex buildings and ar-

cheological excavations. With a CCD-camera (variable focal length)

also integrated in the system, it is possible to record panorama or

detail images that are then available for documenting the scanned

object. Due to simultaneously taking photos while scanning the

scene, true color RGB-values can be assigned to the geometric

XYZdata. By referring to the adjacent points while scanning, normal

directions are determined for the virtual surfaces and are encoded as

Poster

36

so-called compass-colors at each point in the cloud. As these

scanners can only record discrete data sets (point clouds), it is

necessary to mesh these sets for getting surface models.

The meshing process is a complex issue and in the last years a lot of

algorithms were developed to solve this problem. We present an ap-

proximating algorithm for data reduction, which is based on a quadri-

lateral grid. Thus we are able to employ a continuous Level-of-Detail

(LOD) algorithm, from which a simplified 3D surface model can be

created. We start with generating a height map by projecting the

point data onto a plane orthogonal to the color-coded normal field.

The grade of simplification is determined by an error tolerance based

on a measure similar to the Hausdorffdistance. This algorithm is not

a dynamic (view-dependent) LOD mesh simplification, but a nonre-

dundant approximation of point clouds by surface patches. It is a

simple and extendable meshing algorithm, which is made up of

techniques adapted from popular terrain-LOD algorithms. Cracks and

T-junctions are eliminated by merge and split operations. Starting

from a regular meshing of an arbitrary rectangular cutout of point data

a textured surface model can be generated. The resulting model is

composed of a minimal number of triangles for a user defined error

tolerance due to a top-down subdivision algorithm. The presented al-

gorithm is implemented in our OpenGL based editing tool for 3D point

clouds called PointMesh, which will be explained in detail. The sur-

face models then can be exported in VRML-format. For different cutouts

a performance comparison of the regular mesh to various 4-8-meshes

with different error tolerances is given in terms of frame rates.

According to the true color values a texture is automatically produced

on the basis of the finest grid. Coarser grids use the same texture

and appear as complex such that there is almost no loss of informa-

tion during the grid simplification step while the number of triangles

reduces by a factor of five. Automated texturing is possible due to

the color-coded normal field, which allows for a meaningful projec-

tion of curved surfaces onto a plane. This forms the basis for a

screenshot of the texture to be mapped onto this part of the surface

via classical UV-mapping.

We present examples from an excavation site (church rest of Lorsch

Abbey, Germany), the inner part of the Heidelberg Tun (German:

Großes Fass, an extremely large wine vat), and the Old Bridge gate,

Heidelberg.

Session 2

37

HÖRR Ch.

Boon and Bane of High Resolutions in 3D Cultural Heritage

Documentation

Within the past 20 years, optical range scanning devices have

continuously become more accurate, both in terms of lateral and

depth resolution. High-end systems will soon tackle the physical

border that is simply set by the wavelength of visible light. However,

as a matter of fact, increasing resolutions immediately lead to bigger

amounts of data. Hence, even in the view of Moore’s law, it seems

appropriate to weigh up the prospects and problems coming along

with this development. In this paper, we try to figure out, if there is

something like an optimal point cloud density for 3D (and of course

2D) documentation of archaeological finds and which criteria have an

influence thereon.

For the sake of simplicity we will illustrate our arguments using three

fictitious example objects.

• Object 1 shall have a volume of about 1m3, e.g. a head-high

sculpture or column.

• Object 2 shall have a volume of about 1 dm3, e.g. a

mediumsize ceramic vessel.

• Object 3 shall have a volume of about 1 cm3, e.g. a coin or

tooth.

We can roughly estimate the magnitude of the surface area by as-

suming these objects to be a sphere with the denoted volume (1m3,

1 dm3, 1 cm

3). More generally, we refer to one length unit as u.

V =

6

1

πd

3

d = 3

6

π

V

(V

!

= 1[u]3)

d

A = πd

2

A 2

Let us further assume, that three typical optical measuring systems

are available:

• System A shall have a sampling rate of 1 point/mm2

(e. g.

Konica Minolta VI-910 with a field of view of 640x480 mm).

Session 2

38

• System B shall have a sampling rate of 25 points/mm2

(e. g.

Konica Minolta VI-910 with a field of view of 128x96 mm).

• System C shall have a sampling rate of 2500 points/mm2

(e. g. Breuckmann stereoSCAN 3D-HE with a field of view of

48x36 mm).

At first we like to estimate how many single scans are necessary to

capture the whole object. To this end, it may not be sufficient to

calculate the ratio of surface area and field of view, since due its

shape the object may not take the whole measuring area and

therefore several round scans might become necessary. Insofar, the

minimum number of partial scans can be estimated by 10 to 12

independent from the object’s shape. However, the greater the ratio

of object size and field of view becomes the more the tiling effect

comes to the fore. Additionally, we have to assume a realistic

overlapping rate of 50 if not 100% in order to perform a robust

registration [McPherron et al. 2009]. The following table shows a

rough estimation of the number of necessary scans.

Object 1

(4,8 cm2)

Object 2

(4,8 cm2)

Object 3

(4,8 cm2)

System A (3000 cm2) 32 > 10 > 10

System B (120 cm2) 800 > 10 > 10

System C (17 cm2) 6000 60 > 10

While the tasks A1 and C2 can still be handled and realized within

one day, B1 and especially C1 are however questionable simply for

economical reasons. As denoted in the next table, the amount of

acquired data is another big challenge, because after registration

and merging the arising point clouds may become very big.

Object 1 Object 2 Object 3

System A 4,800,000 48,000 480

System B 120,000,000 1,200,000 12,000

System C 12,000,000 000 120,000,000 1,200,000

It can be seen that the number of vertices varies within several

orders of magnitude depending on the object size and acquisition

system. If we assume a memory demand of at least 44 bytes per

vertex, we get the following:

Session 2

39

Object 1 Object 2 Object 3

System A 202.9 MB 2.03 MB 20.8 kB

System B 4.95 GB 50.7 MB 519.5 kB

System C 495.4 GB 4.95 GB 50.7 MB

While the combinations B2 and C3 are already above average, A1 is

a high-density model even from today’s scales. In order to process

4.95 GB of data (B1 and C2) an up-to-date 64 bit system is required

and serious difficulties concerning interactive visualization arise. A

data amount of 500 GB for scenario C1 seems to be an utopic one

even for the near future.

Independent from the limits set by current hardware, the object’s

sampling rate should also be limited with regard to the intended

visualization purpose. Here, essentially four scenarios are

conceivable:

1. The visualization shall be performed on a screen or via beamer.

Even on current HD-ready devices the 2 megapixel border is not

crossed. Consequently, no information gain is achieved above a

vertex count of about 4,000,000, unless highly magnified details

shall be shown.

2. A big poster or an oversize canvas shall be printed. This is mostly

done for popular scientific or marketing purposes for which in most

cases a photograph would have been sufficient as well. Without

doubt, macroscopic properties shall be highlighted and the viewing

distance will in general be several meters.

3. An image in a catalogue of findings shall be generated for

scientific purposes [Hörr et al. 2008]. Usually a scale between 1 : 1

and 1 : 4, for bigger objects sometimes 1 : 10 or even more is

chosen. Plastic details are clearly visible in a normal viewing distance

of 30 to 50 cm.

4. An especially filigree object or a detail view of the object’s surface

shall be magnified true to scale in order to better depict mesoscopic

properties that would not or only hardly have been visible in the

1 : 1 view.

Of course, this enumeration could have been proceeded towards the

microscopic scale, but then a totally different question would be

present that with purely optical measuring systems could not be

processed anyway. In this case a sample scan instead of scanning

the whole surface would be much more convenient.

Session 2

40

Which sampling rate is necessary for pictorial object documentation

(i.e. scenarios 3 and 4), finally depends on two parameters: the

resolution of the printing hardware and the resolution of the human

eye. Current ink jet and laser printers achieve 300 dpi across the

board. Although technically much higher resolutions would be

feasible as well, these are often realized only in high-quality print

media. This is mainly due to the fact that the resolving power of the

human photoreceptors is restricted to approximately one arc minute

even under optimal conditions. Hence, for a distance of 50 cm two

points can barely be distinguished if they are 0.15mm apart

(corresponding to a resolution of 170 dpi). For a distance of 3

meters this value is still 0.9mm (28 dpi).

So if r is the image resolution in dpi and s is the image’s scale, the

necessary (and sufficient!) point distance in object space is

d=1/(r·s). In this case every sample corresponds to at most one

pixel. For a typical resolution of 300 dpi and a scale of 1 : 3 this

would require for example a point distance of about 0.25mm

(acquisition system B). Finally, it should be mentioned however, that

using interpolatory shading techniques such as Gouraud or Phong

shading good results can be obtained even for lower resolutions.

Session 2

41

JUNGBLUT D., KARL St., MARA H., KRÖMKER S., WITTUM G.

Automated GPU-based Surface Morphology Reconstruction

of Volume Data for Archeology

Motivated by new demands of the vast area of ceramics classification

in archeology, we are proposing a novel method to segment volume

data of ancient artifacts using their materials density. This kind of

data can be acquired using industrial Computer Tomography (CT).

Thanks to decreasing costs, CT becomes a more and more reason-

able tool for archaeological surveys. Therefore objects can be in-

vestigated without damage – even without removing the transport

packaging. As segmentation methods of volume data are available in

industry and medicine, we show the adaption of a novel high-per-

formance analysis and visualization method using parallel computing.

The so-called Neuron Reconstruction Algorithm (NeuRA) was

originally designed for reconstructing the surface morphology from

three dimensional images containing neuron cells or networks from

neuron cells. Fortunately, NeuRA also provides the fully automatic

generation of triangular surface meshes from computer tomography

image stacks of archaeology data, like ceramic vases and other

pottery. Figure 1a shows a mesh of a reconstructed ceramic. NeuRA

uses a sophisticated combination of noise reducing and structure

enhancing filters, as well as segmentation methods, a marching

cubes mesh generator and mesh optimization methods. The output

are triangular meshes in different resolution levels. To reconstruct

data sets of several hundreds of megabytes within a few minutes, a

highly parallelized implementation for Nvidia Tesla high performance

computing processors using the supplied Compute Unified Device

Architecture (CUDA) programming library was developed.

Real time rendering of the generated triangular meshes enables

interactive viewing of ceramics. Since the processed data generally

exceeds the device memory of state-of-the art graphic processing

units (GPUs), applying volume rendering techniques is still a

challenge. Performing high data throughputs enables the systematic

reconstruction of arbitrary items. Another benefit is the automated

segmentation of density, which allows to isolate different areas of

interest used as features for archaeological classification. Figure 1b

shows ancient applications of bronze-scales, which are a charac-

Session 2

42

teristic feature of the Este-pottery. The Figure also shows a hidden

metal part supporting the structural integrity of the object.

Figure 1. Reconstructed triangular mesh from a computer tomography scan of a (a)

ceramic vessel and (b) ancient applications (spheres) and a and a metal pin of an old

restauration inside (disc shape).

Presently, the reconstructed meshes only contain information about

the vertex positions and the triangulation. Consequently, future re-

search will include the automatic attachment of textures to the

meshes to provide additional features, like surface colours. Having

widely used 3D-meshes, databases can be established to exchange

the reconstructions among scientists around the world; virtual

exhibitions in museums can be created or access via the world wide

web can be provided.

Session 6

43

KLEPO V., PASKALEVA G.

Artifact Cataloging System as a Rational Translator from

Unstructured Textual Input into Multi-dimensional Vector

Data

Remarkable discoveries are made even at the beginning of the

process of data analysis. Especially by simply trying to sort the

gathered information fragments, not even referring to the more

advanced level of sorting the artifacts. There can be countless ways

in which the information can be organized, but there are only a few

possibilities for sorting the artifacts, in order for them to reveal the

most probable conclusion. The main difficulty lies in fining them.

During my work on the doctoral thesis “Architectural History of Medi-

terranean Lighthouses” I have come across a number of artifacts

which suggested the past of certain buildings might have included

them functioning as lighthouses. Precisely this broad research area

that encompasses the history of so many buildings (in the

Mediterranean coastal region), which are instrumental in the

analytical identification of lighthouses, is a good starting point for

this small scientific research. In this cluster of interrelated artefacts

and information fragments the use of computer technology is of an

essential importance. The beginning of any research activity lies

within the defining of goals, but the true explorer would start one

step ahead. That adventurer would begin with defining the object to

be explored itself. In this case it is the lighthouse. That definition will

be used as the basis of a translating agent.

Due to constructive similarities lighthouses share basic definition

characteristics with towers. The distinguishing features of light-

houses are the specific functions which they had to meet. For

example navigation and signalization have determined the topo-

graphic locations of these structures. So if one should observe more

closely, the idea of the lighthouse does not rest in the architectural

form of the tower, but in the use for which the structure was built.

In the course of history those functions lead a multitude of

architectural types to develop into the form of the tower.

After this introduction, the meaning of the concept of a lighthouse is

easier to grasp. One way in which the historical artifacts concerning

the architectural history of Mediterranean lighthouses could be

Session 6

44

organized, would be to simply follow the chronological order of their

creation. However, such one-sided catalogue would be of no further

use as the development would end with the last added historical

artifact. This thesis suggests an interactive and creative artifact

database. Exactly here the previous definition of the concept of a

lighthouse provides the needed inspiration.

The cataloging should not be based on one or more static criteria.

Instead it should be based on states containing such information as

time, architectural form and uses. These basic elements could be

complemented by building materials, locations (different parts of the

Mediterranean coastal region), whether the structure is part of the

port urban planning or stands alone, etc If these states are then

subjected to transformations by actions designed to achieve a

classification goal based on a few chosen parameters, they could

help formulate new dynamic hypotheses concerning, among others,

the most probable locations of the sites where the lighthouses

existed in the past. The newly organized evidence can be then

translated into, for example, a probable appearance of the

lighthouse under study.

For a person with significant scientific experience in the field of

historical development of cultural heritage most of the historical

artifacts are easily analyzed through named categories and implicit

and even subconscious associations. One of the most significant

tasks of the translating agent would be to transfer the implicit into

explicitly formulated grammar rules of a multi-dimensional recon-

structive language. Thus the more tedious information processing

tasks can be transferred to a computer while leaving the expert and

the interested layperson the opportunity to be as creative as the

human imagination allows. The translating agent can be easily

manipulated to operate on a diverse set of category definitions and

an ever expanding set of language elements, including two- and

three- dimensional vector representation of the gathered artifacts.

This small pilot project attempts to show a way in which artificial

intelligence can be used to satisfy an ever growing need for visual

presentation and accessibility in the field of cultural heritage project

development. It is intended to facilitate decision-making and expert

analysis in order to motivate further work in the area.

Session 4

45

KOR S., BOU V., PHAL D., NGUONPHAN P., WINCKLER M. J.

Practical Experiences with a Low-Cost Laser Scanner

3D laser scanner has become more and more important in the field

of archeology and in other fields in humanities worldwide. In

Cambodia, 3D scanning is being needed to create a virtual museum

for the National Museum of Cambodia in which a large number of

artifacts were not able to present to visitors because of limited area

for exhibition. Above all, an actual preservation and conservation

project of the Banteay Chhmar temple in the northern Cambodia

funded by the Global Heritage Fund, is focusing on scanning temple

stone blocks in order to simulate the reconstruction and

reassembling the stones virtually.

In order to achieve this goal, 3D laser scanner are demanded in

order to reconstruct or record the artifacts and the stone blocks of

the temple. With this project we mainly want to contribute the

Banteay Chhmar Project by supporting some of our first experiences

in 3D laser scanning of the temple stone blocks. We want to discuss

the usefulness of a low-cost 3D laser scanner called David sponsored

by the Interdisciplinary Center for Scientific Computing (IWR) –

University of Heidelberg.

Our team in the Information Technology Center (ITC), Royal Uni-

versity of Phnom Penh, has earlier been awarded a research stipend,

also sponsored by the IWR, on “Investigation of a 3D scanner under

real-life conditions”. With this research, we have scanned a wide

range of objects in completely different environment.

Poster

46

MEIER Th., CASSELMANN C., VAN DE LÖCHT J., KUCH N., ALTENBACH H.

A new approach to the surveying of archaeological

monuments in academic teaching

A main issue concerning application-oriented academic teaching in

archaeology is the training in different excavation techniques. To

mediate this knowledge, the Institut für Ur- und Frühgeschichte und

Vorderasiatische Archäologie at the Ruprecht-Karls-Universität

Heidelberg offers training excavations. Here, students learn how to

deal with the methods and techniques needed in archaeological field

research in a practical setting.

The academic training of archaeologists also includes a basic

knowledge of geodesy. Hence, courses in surveying methods are

offered for students of all archaeological subjects. This training is

divided in three different modules. In the first, the students acquire

a basic theoretical knowledge of geodesy. In the second, the

students use this theoretical knowledge in different practical

exercises. The aim of the third is to broaden the basic knowledge

won in the first two modules. In this hands-on seminar, an

archaeological monument is measured with an electro-optical

theodolite also known as a total station. Afterwards, the data is

edited into a complete mapping using a Computer Aided Design

(CAD) application.

Since the summer of 2009, a Topcon Imaging Station (IS) is avail-

able to our institute. This new total station serves for conventional

tachymetry as well as for laser scanning and photogrammetry. The

software (ImageMaster Pro) allows the rendering of stereoscopic

pictures of the measured monument during the post-processing in

texturised triangular polygons. The first measured monument was a

medieval castle situated in a cave in the Luegsteinwand near Ober-

audorf (Rosenheim district, Upper Bavaria).

West of the village Oberaudorf, there is the Luegsteinwand, a face

that raises perpendicularly up to a hight of 300 meters above the

valley floor. The measured cave with a width of approx. 14 meters

and a height of nearly 7 meters lies almost in the middle of the face,

extending nearly 25 meters into the rock. Visitors may reach the

cave by means a steep path and a rope as well as a ladder with a

height of 6 meters. The first excavations took place in the years of

Poster

47

1967 and 1968 and were led by the local pastor. In 2008, Thomas

Meier led the first professional archaeological excavation, during

which this a floor plan was drawn using a total station. Since it would

cost considerable effort to geo-reference the site into the Gauss-

Krüger coordinate system, we used a local coordinate system.

The cave yielded mostly medieval pottery dated between the 11th

and 13th century. Greyware, which is typical for the late medieval

times and the renaissance, is totally missing. Moreover, some

fragments of historical and early medieval pottery occurred. The

most impressive of the few preserved building remains is the front

wall of the cave, which reaches over 5 meters in height with a heavy

thickness of 1.3 meters. This wall sealed the cave sand prevented

erosion of the archaeological layers. 14

C-samples of the pits

belonging to the cave dated the 11th and the beginning of the 12

th

century which correspond with the pottery findings.

Based on the results of the survey of 2008, in the summer of 2009

extended the surveying. This time, a three-dimensional measure-

ment was made possible by using the new IS. Carsten Casselmann

led the survey within the Vermessungskunde III seminar mentioned

above. The individual structures were measured from different

positions using the IS contact-free laser sampling over the whole

surface of the cave. The density of the measured points was

adjusted to the structure as well as to the different positions of the

total station that determined the measurement angle. Two days of

measuring produced a cloud consisting of more than 500,000 points.

The irregularities of the cave wall really put the instrument and

engineer to the test. In a post-recording step on the computer these

points were meshed into triangular polygons and overlayed with a

texture skin. Hidden edges or naturally cavities remained minor un-

solved problems. The vegetation posed a further problem, because

the IS could not discriminate the ceiling and the walls from the

plants. We edited the vegetation out of the ground, but this proved

impossible in the time alotted for the walls and ceiling. Thus, the IS

could not differ between the plants hanging the roof and the

stalactites. Owing to weak illumination, the photos made auto-

matically by the IS were useless. Instead we constructed an artificial

texture in the reconstruction.

Despite the difficulties, in all the results of the measuring was

adequate. In other such laser measurements, such as scans of

Poster

48

architecture or archaeological excavations, such extreme conditions

need not be a problem.

The models created will be integrated in a new trail, which is part of

the local tourist attraction program sponsored be the EU-Leader+

Program.

3-D-View of the cave from E to W 3-D-View of the cave from W to E

Session 1

49

NEMES P., GORDAN M., VLAICU A.

Color Restoration in Cultural Heritage Images Using Support

Vector Machines

Introduction

Color restoration in digital images is necessary because of physical

degradation and of imperfections generated during the acquisition or

visualization process. The color restoration process implies either a

physical model of the deterioration process, or the derivation of the

correction function from examples (in case an estimation is difficult

to make – often the case in practice). Research is done in both

directions; but, in the second case, there is the advantage of a more

flexible method, if one assumes the same degradation conditions. In

this case we can have highly nonlinear correction functions. The

color restoration function can be defined as transformation of a

“deformed” color space into a “correct”, ideal, space. It can be

defined either as a set of three scalar functions, component-wise, or

as a vector function, if the three color components are correlated.

This paper aims at examining the efficiency of supervised learning

methods in the derivation of color correction functions. From the

existing supervised learning techniques, some of the most appealing

in this field are: neural networks; support vector machines (SVMs)

used for regression. In the last decade, SVMs are especially used on

a large scale in classification and regression, but their use in color

restoration of digital paintings, affected by various ill-defined types

of degradations, is still limited; however, as research shows, they

can be a promising alternative to other restoration methods. That is

why we focused on their use for color restoration of degraded

paintings, examining their performance as compared to the experts’

restoration.

The proposed approach to color restoration in digital paintings by

support vector regression

SVMs are based on a powerful learning paradigm based on the

structural risk minimization. The ability of SVMs to learn from a

relatively sparse and reduced set of training data with good

generalization and recall performance could make them an excellent

alternative to other machine learning methods for color restoration in

Session 1

50

digital paintings – which is a mathematically hard to define problem,

and where often the set of training data can be considered sparse

(the set of training samples in the form of degraded – physically

restored image is limited). Initially defined as binary classifiers, the

SVMs were extended for the regression issue (SVR – Support Vector

Regression). Starting from a pair of vector input – scalar output

training data set (typically denoted by (x,y), with x – the input

vector in ℜN

, N y – the real valued scalar output), SVM

learns the functional dependency between the possible values of x

and y, in the form of a function f(x): ℜN

→ ℜ .

Let us consider a color deteriorated image (preferably one with a

large color spectrum in order to cover as much of the color space as

possible) and let us assume that the restored (desired) image is

available (in the case of degradation of the paintings physical

support – the image restored by experts). For restoring any other

image degraded in the same manner, one must identify the

transformation function of the color space (in vector form or on each

component, if the color components are decorrelated) – which can

be provided by SVR. Since the default SVR result f is not vector

valued but real valued, the color vector mapping cannot be obtained

directly; instead we derive three color mapping functions, one for

each color space component, using a rather decorrelated color

space, as e.g. YCbCr. The best performance was experimentally

obtained using scalar input – scalar output mappings by three SVRs

on each component individually, denoted by fY(Y), fCb(Cb), fCr(Cr)),

and non-linear SVRs (especially with Gaussian RBF kernels). This

restoration scheme is illustrated in Fig. 1.

Fig 1. The support vector regression based restoration scheme

The training set of the three SVRs was generated by selecting a

representative pair of corresponding patches depicted from a

degraded and a restored painting (spatially aligned at pixel level).

Session 1

51

The set of values of the three color components from the given

degraded image and from the restored image were used to generate

the three training sets (one per color component: Y, Cb, Cr). After

training three SVRs on each set, the resulting mappings fY(Y),

fCb(Cb), fCr(Cr)) are applied to the each color in the degraded

painting, to obtain the color restored painting accordingly.

Implementation and results

The implementation of the proposed SVR based color restoration

scheme was done in Matlab, using a publicly available SVM toolbox.

For training and testing, some degraded and restored painting

images provided by K. Nicolaus (1999) were used; the patches

depicted to generate the training data were approximately 20×30

pixels. Several SVM configurations were examined, with different

kernels: linear, polynomial, Gaussian RBF and exponential RBF. The

Gaussian RBF kernel SVR provided the best performance in respect

to the error and to the visual similarity between the SVR and expert

restored image. An example of a degraded painting, its expert

restored version and the result of the proposed SVR based

restoration is presented in Fig. 3. The training patches used for the

restoration of the image in Fig. 3.a) are shown in Fig. 2.

Fig. 2 Training images: a) input image (degraded), b) output image (physically

restored); the results of the restoration process applied on the input training image

using the regression functions applied simultaneously on Y, Cb and Cr with: c) liniar

kernel; d) polynomial kernel; e) gaussian RBF kernel

Experiments were run on four digital paintings from the same

source, with different color palettes and degradation levels. As

expected, the color restoration performance decreases dramatically

for very faded colors. In numerical terms, the average error between

the SVR color restoration result and the expert restoration result

over the test set decreases from 10.8% in the linear SVM case to

4.1% in the case of Gaussian RBF kernel based SVM.

Session 1

52

Fig. 3. An example result, showing: the degraded image (a); the restored images

obtained by applying the regression functions on the Y, Cb, Cr channels of the

degraded image (b) through d)), using an SVM with: b) linear kernel; c) polynomial

kernel of degree 2; d) gaussian RBF kernel. The last image (e) shows the result of the

physical (expert) restoration.

Conclusion

The work presented here proposes a simple nonlinear color re-

storation approach for digital paintings, based on support vector

regression in the YCbCr color space. The experimental results show

that SVR application on cultural heritage image restoration has great

potential and further approaches for color restoration based on SVR

worth investigated. Future work should address the issues of finding

the optimal color space, selecting optimally the training set, and to

derive more advanced SVR architectures that could optimize the

color restoration performance in cultural heritage images–able to

adapt to the specific painting palette and degradation.

Poster

53

NICCOLUCCI F., NYS K., FELICETTI A., HERMON S.

The Hala Sultan Tekke site and its innovative documentation

system

Hala Sultan Tekke (HST) is a Late Bronze Age site located near

Larnaca, Cyprus. The site was explored by two British expeditions at

the end of the 19th century, yielding a rich set of objects of gold,

silver, bronze, faience, ivory and pottery. Some tombs were investi-

gated by the Cypriot Department of Antiquities in the ’60s, under the

direction of Prof. Karageorghis. Extensive research on the site has

been conducted by the Swedish Prof. Paul Åström since 1971 until

his death in 2008, with the publication of an impressive corpus of

reports. Since 2001, Prof. Karin Nys has been the Assistant Director

and currently she in charge of all the archaeological research on the

site, in particular of the publication of the results of the most recent

campaigns directed by Prof Åström that remained unpublished due

to his death. Records consist of notes taken by him and the trench

masters using “forms” for finds and features, which were filled rather

freely with their observations, frequently regardless of the headings

present in the form and almost always handwritten. They also

include photos, drawings, plans and maps. Notes are more similar to

free text than to structured forms, and geographical and spatial

information is very often included.

A simplistic solution trying to summarize the notes into a structured

repository such as a DBMS would filter data accepting what is

deemed as “relevant” and forcing it into the grid of the data

structure, arbitrarily superimposed on the information recorded on

the field. Such a solution would prevent distinguishing the selection

operated now from the original archaeological record. On the other

hand, recording the notes just as simple text would reduce queries

to free-text search only, and exclude any possibility of geo-

referencing. An innovative solution must then be implemented.

Therefore two design decision were taken:

1) Use a markup solution for the notes, annotating the text deriving

from their transcription basing on a CIDOC-CRM compliant ontology,

preserving the richness of the text, enabling semantic queries and

basing markup on a standard.

Poster

54

2) Create a system based on Open Source software, enabling geo-

referencing of such elements in a true GIS system.

The first step consisted in defining a task ontology. The HST

ontology is a subset of CIDOC-CRM, with the geographic extension

incorporating GML, the Geographic Markup Language. The HST

ontology includes all the semantic elements used in the records.

However, if additional elements will be necessary in the future, the

annotation procedure will easily incorporate such new elements from

then on. Although notes appear to use a consistent terminology, the

system includes a thesaurus to associate preferred terms, and a

gazetteer listing infra-site location names or codes currently use, for

easy and standardized reference.

An annotation tool was created for data input. It allows to markup

text while typing it or to apply markup to already typed text,

selecting the relevant text and choosing the appropriate element

from a list, storing this information in an efficient retrieval system. A

similar procedure takes place for properties, i.e. relations among

marked elements, providing a way to resolve cross- and co-

references. Through these tools, a conceptual semantic layer is

superimposed on top of the base text, i.e. the original records

(actually, their faithful transcription), which will remain untouched.

The association to designated areas in the GIS works in a similar

way. Areas, lines or points to be referred need prior definition –

possibly interspersed with text markup, temporarily activating the

GIS module when necessary. Then they can be associated to text

when they are referenced in the text records, implicitly or explicitly.

The system includes several modules: the already mentioned one for

text encoding (Annotation Tool), a second one for searching the

repository (Query Interface) and modules for interfacing the GIS.

As GIS engine, we chose QuantumGIS, a multi-platform open-source

GIS, for the flexibility of its modular architecture of core functions

and plug-ins, allowing a high degree of personalization and

integration with other frameworks.

The resulting system allows any complex query, but to facilitate non-

specialist users some have been incorporated in the Query Interface.

This includes: a simple text query builder, relying on the thesaurus

and gazetteer; a semantic query builder, using all the classes and

properties provided by the HST Ontology and by CIDOC-CRM,

optionally refined by adding a string for domain or range entities;

Poster

55

and a “faceted” browser, i. e. for searching the repository by

progressively narrowing the scope of the quest.

Since geographic information is stored as GML, these operations

include also geographic queries. On the other hand, the GIS may be

used to visualize the outcome of a query.

So far, the system has been tested on a subset of data. The

completion of data input will require rather a long time for the large

amount of records to be transcribed, while encoding will not delay it

substantially.

If the system will work as expected and as confirmed by tests, it will

be a general-purpose tool useful in all those cases in which the

source is a text, for example an excavation diary or a report. It may

be applied to historical sources as well, and mix different and rather

inhomogeneous sources – even with different underlying ontologies,

as CIDOC-CRM provides the semantic glue to join them. Inclusion of

GIS features adds a spatial dimension, which is paramount not only

for archaeology. In a similar spirit, the general system in the future

may include any external viewer, for example for 3D objects, as long

as the related ontology is incorporated into CIDOC-CRM and the

package allows customization to add the necessary plug-ins.

Additional plug-ins may finally allow heterogeneous query

combinations, for example mixing pattern recognition with text-

based queries.

Poster

56

PECCHIOLI L., CARROZZINO M., MOHAMED F.

ISEE: retrieve information in Cultural Heritage navigating in

3D environment

Access to information enables knowledge to be shared among

people; it is therefore becoming increasingly important. Moreover

our notion of space is changing, as is the way we see it.

Cultural heritage is an interdisciplinary field that draws together

several different professions. Information is gained from different

sources and in varying formats. Furthermore, the relationship

between the conservation managers, who are often unfamiliar with

current documentation techniques, and the providers of the

information, who tend to be highly technical practitioners without

expertise in cultural heritage, is not easy to handle. Moreover, in

Cultural Heritage objects often have a strong 3D component, and

cannot be easily represented with conventional data management

frameworks like Geographic Information System (GIS). The use of a

3D framework may allow a closer adherence to the real world, as it

respects the spatial relationships among various parts.

Starting from these important points, we developed a method called

ISEE (“I see”). It has been the result of a Doctor Europeans in

Technology and Management in Cultural Heritage, “Accessing

Information Navigating in a 3D Interactive Environment”.

This method allows spatial information to be accessed through the

interactive navigation of a synthetic 3D model, reproducing the main

features of a corresponding real environment. The user can get the

pieces of information more relevant where he/she is looking at it.

The system can be used with standard Web browsers, allowing

access to a wider audience without any special requirements1.

The information is naturally embedded in 3D spatial contexts, and

the most natural way of behaving in such a context is to just look

around; indeed, seeing is an essential action common to humans

and in our natural behavior since birth. By using sight we naturally

focus on something (the origin of the name of the method);

therefore, we propose to use this action as a common language

which people can use to understand, query and insert information.

This is the core of the method: using sight to actively query and

insert information. This realizes one of the most important goals in

Poster

57

our approach: to ease everybody’s access to information. The

complexity dictated by the type of data is simplified in a few easy

moves (touching a model in a device and looking around); moreover,

we think that this involves also an amusement component which

helps users in getting more involved in the experience (Fig.)

Fig.: the last version of ISEE

The use of extended zones gives to the proposed ranking algorithm

a superior performance than rankings based only the distance or

selection methods. The system has been applied to selected case

studies related both to outdoor and indoor environments, proving

potentially to be also an interesting prototype as a smart guide with

the use of augmented reality technologies.

We extend our previous work on the ISEE method to link information

to 3D space to a mobile platform.

Our goal is to use it in a mixed reality setting where the user can

retrieve information about his/her current context using the GPS

information given by the device for the outdoor application, and

marker-based mixed reality for indoor environments.

The actual prototype is focused on an application in Berlin using a

smart device. The approach predicts a platform for a common device

provided of graphic interface to visualize Google Maps, or of the

performances to receive the position and so to visualize the list of

the information. In particular we are working already analyzing

performances of the Iphone 3G. The relatively low accuracy of the

Poster

58

localization information has to be kept into account in the design of

the interface.

We allow the user to correct his location and explore the surrounding

of the location returned by the GPS freely, while continuously

integrating the changes of position given by the GPS. The user has

access to the relevant information for the current context, in a

similar way as done previously in the ISEE Web application, but with

a slightly modified interface to take advantage for example of the

touch screen of the device, and to keep into account the smaller size

of the display. The data itself is gathered and stored using a REST

interface to the ISEE Web server.

The goal of work is walking to retrieve a contextual information in

real time by an intuitive interaction. Information insertion is simple,

as the user can add in every moment specific data about his current

view. We hope in the future to let normal users add information,

fostering the creation of a community of active users able to deliver

information and share experiences.

Session 3

59

PEREIRA J., STRASSER A., STRASSER M., STRASSER Th.

Interactive Narratives for Exploring the Historical City of

Salzburg

This paper presents the main findings of the “Wanderbarer

Salzburgführer”, an interactive, mobile explorer for the historical city

of Salzburg. The city explorer was developed as part of the

INTERREG IIIB CADSES project Heritage Alive!, and aimed at

exploring and testing interactive narrative approaches in presenting

cultural heritage content to the residents of Salzburg. In particular

we wanted to test how interactive storytelling techniques can

improve the local residents’ experience in exploring their own city’s

cultural heritage. The city explorer presented six subjects, which

focused on less-prominent heritage of Salzburg:

• Salzburg Through the Ages;

• Georg Trakl: The Life and Works of a Salzburg Poet;

• Historical Taverns, Breweries and Hotels;

• Latin Inscriptions: What Inscriptions Tell Us;

• Historical Windows;

• Historical Doors.

The city explorer adopted a moderate constructivist narrative

approach. This approach includes elements of constructivism (i. e.

users actively construct an understanding of the world through

personal interests, experiences and notions) and elements of

instructive design to provide users with structured guidance and

support for creating knowledge. Key aspects of constructivist

learning include encouraging users to participate in active learning,

authentic learning and to provide them with multiple perspectives on

knowledge creation. To provide users with structured guidance and

support we followed the SOI model developed by Richard E. Mayer.

This model aims at fostering the cognitive processes of users in

knowledge construction in terms of selecting relevant information

(S), organizing information in a relevant way (O) and integrating

new information with users’ prior knowledge (I).

The interactive storytelling techniques applied included various

interactive features as well as a dynamic user navigation system.

The interactive features included quizzes, scavenger

hunts/geocaching, answering of single/multiple choice questions,

Session 3

60

various games and riddles and hyperlinks for accessing additional

content. For example, users were able to complete a quiz to learn

more about the history of Salzburg’s main bridges; the Historical

Windows and Historical Doors themes followed a scavenger hunt

approach (users were provided with an image of a historical door or

window and the approximate location of where to find them). In

another theme, various historical taverns, breweries and hotels were

introduced through a brief history and anecdotes that highlighted

funny incidents, oddities and other memorable details. Users could

also learn more about the numerous Latin inscriptions found all over

the historical city of Salzburg: As most users could not interpret

them, they were provided with guidance to learn about their

meaning and their historical context. Users were also encouraged to

create content and share their thoughts and ideas with other users.

For example, users were invited to complete or comment on selected

poems by Georg Trakl.

Another major element in extending the user experience was the

ability of users to dynamically create their own paths through the

city. Using the mobile explorer’s dynamic navigation system, users

were able to visit places of most interest, or establish links between

different themes—i.e. they could visit places nearby, or thematically

or historically related to their current point of interest.

The “Wanderbarer Salzburgführer” mobile application is built on a

multi-tier Enterprise Content Management system (ECMS, silbergrau

blueContent). As a web browser application it is client independent

and can easily be adapted to different end user devices, such as

mobile phones, Playstation Portables or UMPCs. In fact, three

different user interfaces were implemented: one aimed at mobile

clients (netbooks, UMPC), one aimed at web browsers and one

specifically for a mobile device (Blackberry). Technically, all are

based on XHTML and JavaScript.

Various engines were built on top of the ECMS to provide the

following functionalities:

• Story Engine for creating stories in a linear manner;

• Quizzes and Riddles Engine to create and integrate quizzes

and riddles into the stories;

• Location Engine to locate users at any time during their

tours through the city (using GPS technology and Google

Maps); and

Session 3

61

• Multi-relation Engine to provide the overall dynamic

navigation and link all digital content.

The mobile city explorer was trialled on a Medion UMPC by 15 test

users in the summer of 2007. Test users (the younger users in

particular) took advantage of the interactive features and dynamic

navigation to explore the City of Salzburg. They commented that

with the ability to switch views and themes effortlessly they were

able to quickly select their own path through the city. And they also

remarked, that with activities, such as searching for historical

windows and doors, or Latin inscriptions their experience of the

objects was more intense as though they were rediscovering the city

through “different eyes”.

Session 1

62

QUATEMBER U., THUSWALDNER B., KALASEK R., BATHOW C., BREUCKMANNB.

The Virtual and Physical Reconstruction of the Octagon and

Hadrian´s Temple in Ephesus

The Octagon is a tomb-monument situated at the western end of the

Curetes Street in the centre of the historic site of Ephesus and was

erected in the first century B.C. It was excavated in the early 20th

century under the direction of R. Heberdey. At that time the building

was regarded as a kind of trophy-monument. In 1926 M. Theuer

started an excavation on the top of the base of the structure and

discovered a barrel vault that contained a burial chamber. Inside,

they found a sarcophagus with the skeleton of a young woman.

According to the interpretation of H. Thür, there is sufficient

evidence that the unknown woman inside the sarcophagus is

Ptolemy Arsinoe IV, the youngest sister of the famous Cleopatra VII.

During the excavation of the base structure and the area of the

lower Curetes Street, numerous blocks of the ruined building were

found as well. Therefore W. Wilberg, who was the architect of the

original excavation, could already provide an initial reconstruction

plan of the building that does not differ very much from the current

one.

The Octagon was built on a quadratic base. In total, it is 13 meters

high and its front is subdivided into three parts: a pedestal, above

this the octagonal main structure - to which it owes its current name

“Octagon” - and finally a

steep pyramid-shaped roof.

Due to its historical

relevance and its prominent

location in the centre of the

excavation site next to the

Hadrian´s Temple and the

Library of Celsus there is an

increasing desire to rebuild

this fascinating building at

its original location. Most of

the fragments are currently

situated on the site, but two

columns and some parts of the cornice were transferred to Austria in

Session 1

63

the early 20th century and are now on display in the Ephesus

museum in Vienna. Modern 3D technologies provide the means to

digitally reassemble all of these fragments that are physically located

in different places.

For this reason the current project focuses on working out an

anastylosis in virtual space with the aid of highly detailed 3D-models

of all remaining components of the building. All building parts were

recorded computationally with help of 3D scanning technology.

Two different 3D-scanning systems were employed for data

acquisition, a time-of-flight laser scanning system for the entire en-

semble, and a structured-light triangulation scanner for more de-

tailed structures. From this data 3D-models of the remaining blocks

were generated which formed the basis to create a virtual ana-

stylosis of the entire building.

The so-called Hadrian’s

Temple with a size of

approx. 10 m x 10 m and

a height of about 8 m is

one of the most famous

monuments in the ancient

city of Ephesus and oc-

cupies a prominent lo-

cation in the western sec-

tion of Curetes Street,

one of the chief thorough-

fares of the site. Although

it was discovered in 1956,

this structure has never

been systematically analyzed, studied, or published.

As a result, it has remained a subject of controversy for over half a

century. Until now, scholars have been unable to ascertain its

chronology, function, and definitive architectural reconstruction. A

new project conducted at the Austrian Archaeological Institute and

funded by the Austrian Science Fund (FWF Project P20947-G02) is

designed to redress all these problems.

In the preliminary excavation report, Franz Miltner, the then ex-

cavation director, interpreted the building as an imperial cult temple

for the emperor Hadrian (117–138). However, this interpretation is

contradicted by a subsequent study of the building inscription. So

Session 1

64

far, a consensus concerning the function of the building has yet to

be established.

In the summer of 1957, the actual reconstruction work began. The

short period of time between discovery and the reconstruction of the

building left little time to study building history and phases. Until

now, only a plan and a restored elevation have been published.

Documentation of the architectural blocks is not available in the

archives of the Austrian Archaeological Institute. Before issues such

as interpretation and function can be addressed, it is of vital

importance to clarify the architectural history of the structure. The

first step towards this goal is to produce an up-to-date

documentation of the building. This will be achieved by 3D scanning

conducted by the Austrian Archaeological Institute in cooperation

with the Breuckmann GmbH.

The objective of

this project, which

has been carried

out in June/July

2009, is to produce

a high-definition

3D-scan of the

temple with a

resolution of half a

mm by using 3D-

scanners that utilize

fringe projection

technique in combi-

nation with photogrammetry.

The paper will start with a short introduction into the history of these

two buildings. It will then give a short overview about the

techniques, used for this challenging project. The second part will

present the virtual 3D-model of the Octagon and first results of the

scanning and the virtual reconstruction of the Hadrian´s temple.

Nearly 2000 scans have been recorded within 10 nights, most of

them on a scaffolding or crane in a height of about 8 m.

Opening

65

REINDEL M.

Remote Sensing and 3D-Modelling for the Documentation of

Cultural Heritage in Palpa, Peru

In the last ten years our archaeological group had the opportunity to

cooperate with research teams of other disciplines in the framework

of the Nasca-Palpa Archaeological Project, on the south coast of

Peru. Many scientific methods and engineering techniques have been

tested and applied in order to document the cultural manifestations

of the cultures that developed in the Palpa valleys in the northern

Nasca drainage and to reconstruct the settlement history of the

region. In 2002 with support of the German Federal Ministery of

Education and Research we established a project group which aimed

at the development or adaptation of new scientific methods and

technologies for archaeological research. From an archaeological

point of view, the major challenge - but also the major outcome of

the project - has been the focus of the research efforts on one single

region. As a result, a maximum of information could be achieved

about the cultural development during the prehispanic periods of the

Palpa valleys.

In this talk I’ll focus on

different methods of

remote sensing and 3D-

modelling used in the

research project: Land-

scape modelling with

satellite images of dif-

ferent sensors and

aerial photos; site

modelling with photo-

grammetric methods,

laserscanning and tachymetric survey; architectural modelling with

CAD programs; modelling of big objects with laser scanning and

photogrammetry as well as modelling of small objects with laser

scanning and structured light. The applications will show the

advantages and limitations of the different techniques in the specific

application during our archaeological research in Palpa.

Session 5

66

SAUERBIER M.

Image-based techniques in Cultural Heritage modeling

The importance of accurate and visually attractive documentation of

Cultural Heritage, be it single buildings, monuments or whole

landscapes, has grown more and more over the past years.

Compared to other 3D measurement techniques, image-based

methods provide the opportunity to produce photo-realistically

textured 3D models of such objects. Moreover, image-based

methods are quite flexible in terms of accuracy and the level of detail

required in different applications and therefore are more and more

being used in disciplines dealing with documentation but also

analysis of cultural

heritage, in par-

ticular in archaeo-

logy and architec-

ture. Based on

different case

studies accom-

plished in our

group an overview

of applications of

photogrammetric

reconstruction of

cultural heritage

objects will be

given. Furthermore, current topics of research and future trends in

the field of photogrammetry with potential benefit for the Cultural

Heritage community as well as the use of 3D models beyond mere

documentation will be briefly presented and discussed.

Session 5

67

SCHULTES K., BERNER K., GIETZ P., ARNOLD M.

GIS and Cultural Flows

The cultural flows which are researched in the Heidelberg Cluster of

Excellence “Asia and Europe” all exhibit both, a historical and a

spacial dimension. While browsing the factual data they involuntarily

are localized on everybody's mental map–which may be faulty or

incomplete. Therefore, the visualization of the key concepts in a

multidimensional way (geographical position plus timeline) may

foster the understanding of the data included in various database

projects established for the researchers of the project cluster

themselves as well as for the “interested public”.

The project uses Google Earth for which an increasing amount of

additional tools and interfaces are available. It’s main advantage is

its easy-to-use interface: Researchers in the humanities are in most

cases not familiar with specialised GIS-Software. This product is wide

spread and–for the users–available for free.

The Cluster project “Mapping the Flows” is using Google Earth's

"time series" software in two demos. One demo visualises how in-

scriptions were added over time to a Buddhist sutra passage at

mount Tai in Shandong province, China, beginning in the 6th

century.

The other demo is much more complex. It shows the travels of Ger-

man emperor Friedrich II and links each place to the passage of his

itinerary as recorded in the Regesta Imperii on-line database (http://

mdz1.bib-bvb.de/cocoon/regesta-imperii/kapitel/ri05_fic1881_kap10).

While coordinates for the demos were still entered manually we are

now working on a solution that automates this process. This includes

the digitisation and storing in a database of lists of place names

available in the Latin-English “Orbis Latinus”, and geographical

thesauri, like Getty’s Thesaurus of Geographical Names (TGN). It will

also include a tool to automatically assign geo-coordinates to places.

It has shown, that such a visualisation not only makes historical cul-

tural flows very intuitively accessible and understandable. In addi-

tion, this technology also opens new possibilities for validating the

data in respect to accuracy and probability. The paper will introduce

the demos and discuss possible ways of using Google Earth and how

these can provide new insights for the researchers in the under-

standing and analyzing of historical data.

Poster

68

SDYKOV M., SINGATULIN R.

IT of the Project "Virtual Ukek – Kyryk-Oba"

In 2001 on the territory of Uvek (Saratov) the research work was

carried out. The purpose of the researches was to search the traces

of medieval buildings with the help of advanced technologies. As tool

means there were used multispectral stereophotogrammetry

cameras. As a result of researches 81 structures and traces of more

ancient structures were revealed. On thermo graphic "traces" the

sites of stone buildings were identified. "Traces" of buildings were

used for reconstruction of topography of a medieval city. Archaeo-

logical work of 2002–2006 have proved the used decisions. The

impossibility to carry out the further researches in conditions of

modern city, complexity of legal and social character, have induced

to search new basic decisions. It has resulted in intensification of

virtual reconstruction of medieval city, realization of works on safety

of the territory with the help of Web-technologies and creation of an

infrastructure of the historical-tourist complex (the project "Virtual

Ukek", 2005). The correctness of statement of the research problem

was appreciated proceeding from physical-archaeological model of a

monument constructed on a priori information, determining depth of

researches, efficiency of application of various technologies. Taken

as a whole, this data formed the purposeful information bank with

the definite set of the entered data. Use of mathematical methods

(system analysis, the research of operations) has resulted not only in

high scientific results, but also in achieving large technical and

economic effect. The researches have specified an essential role of

information and communication technologies, especially when the

problems of reception of operative advice, performance of functions

of maintenance of security measures and other tasks. Occurred the

successful application of advanced technologies in the project

"Virtual Ukek", has allowed to use them in the extended version at

researches of medieval cities Bulgar and Buljar (the project "Virtual

medieval cities of the Volga region: Ukek-Bulgar-Buljar", 2006), and

also to set off to the research of the territory of the Zolotorevskoe

settlement (Penza, Russian Federation). In 2008 the working group

joins Western Kazakhstan archaeological Centre. In sphere of

scientific interests entered unique royal burial mounds "Kyryk-Oba"

Poster

69

(Fig.) and found in 2001 during the examination of the pipeline route

"Aksai-Atyrtau" the medieval settlement "Zhayik" in the suburb of

Uralsk (Republic Kazakhstan). For research, protection, and

reconstruction earlier proved technology of the project "Virtual Ukek"

are used.

Fig. 1: Reconstruction of the king’s burial mound «Kyryk-Oba»

The features of the used technical decisions on the basis of the Web-

appendices and multispectral stereophotogrammetry technologies

consist in an opportunity of realization of research works in con-

ditions of urbanism, in sharp decrease of cost of shooting, re-

gistration and data processing in a mode of real time. The expenses

for service, support, protection of monuments of a cultural heritage

are reduced. Essential stage of application of technologies is granting

access in virtual space, demonstration of results of application of

technologies. The works are carried out with the help of information-

measuring systems (multispectral stereophotogrammetry cameras)

and combination of methods of the spectral analysis of stereo

snapshots training classifications, construction of the index images,

visual and multispectral decoding etc. The synchronous work of the

Web-appendices and technology imposing threedimensional

archaeological models on modern environmental conditions with

identification of basic attributes (digital thematic databases) gives

opportunities of realization of joint researches in a mode of real time

and transition to ways of simultaneous study of objects with the help

Poster

70

"of a method of immersing". Advantages of offered technology are

based on the reception of the operative information in a mode of

real time as directly from the camera (by means of the protocol of

family IP), and through special video server, and also application of

the developed original software for accelerated photogrammetry of

processing of stereo pairs.

The project stipulates the use of the software complex PHOTOMOD,

which unites an extensive set of software for digital photogram-

metrical processing of the data of distance probe, allowing for

getting spatial information based on the images of practically all

commercially available shooting systems, such as digital and film

cameras, space scanning systems of high solution, use of scanned

archive photographs, etc. Thanks to the flexible module structure,

API for creation one’s own modules of extension (plug-ins),

extensive net of exchange formats PHOTOMOD may be used as a

local fully functional digital photogrammetrical station, a distributive

net media for the realization of big projects and conduction of more

labor-consuming processes. Testing information system with applica-

tion of high-speed networks of transfer of the data now is made. The

separate decision of used technologies is the opportunity of visuali-

zation of the reconstructed three-dimensional image directly on

objects of culture – that the user sees without any special

adaptations. At present the usage of the systems of virtual reality in

the project is one of the most interesting spheres of research and

designs in application of the computer graphics. The system inter-

acts with the images obtained from the multi-spectral photogram-

metrical cameras and with the archeological database and thus

models virtual media. It allows for the creation of the widened reality

with the visual identification of the placement, major indications,

properties and the material of the three-dimensional archeological

objects. The usage of specialized structures of virtual reality is not

obligatory for the project, but it essentially enlarges the effect of

"immersion".

Session 1

71

SIART Ch., EITEL B.

Digital Geoarchaeology – an approach to reconstructing

ancient landscapes at the human-environmental interface

Geospatial tools such as satellite remote sensing, GIS and 3D

visualisation have become more and more popular in modern

archaeology, most notably with regard to the reconstruction of

ancient landscapes and cultural heritage management. Their

strengths and limitations have still not been fully explored, though –

a fact which specifically holds true for the type of applied datasets in

terms of generation, quantity and quality. This aspect is of particular

importance since obtaining and developing useful environmental

data can be the most time consuming and costly aspect of

computerised archaeological projects. As there is a huge demand for

detailed knowledge about the environment, comprehensive data

bases are required in conjunction with proper ecosystematic

information to provide maximum representativeness of results.

The paper aims to bridge the gap between the acquisition of

profound geodata for archaeological issues, their processing and the

appropriate methodological background. Associated potentials and

pitfalls will be focused on from an operator’s perspective as well as a

geographical point of view. Several applications and results from an

interdisciplinary research project are presented. In this context, the

main objective is to reconstruct the Minoan landscape of the island

of Crete and its evolution during the Bronze-Age. Corresponding case

studies were carried out on the basis of satellite imagery, digital

elevation models, GPS data, geomorphological prospection and

literature surveys. As special attention is paid to human-

environmental interactions, the approach highlights the promising

prospects of digital geoarchaeology.

Preliminary investigations focused on reconstructing the socio-

economic and cultural landscape during the 2nd

millennium BC by

using GIS-based algorithms for deriving former roads and

connections between Minoan settlements. Therefore, least-cost

paths were calculated on the basis of GPS-mapped locations and

subsequently compared to communication routes known from

archaeological literature. For the first time, a hypothetical mesoscale

Session 1

72

road network of Central Crete gives an impression of the spatial

interconnections between different Bronze-Age sites.

Besides exploring the ancient landscape by calculating these linear

infrastructures, the main objective was to reveal hitherto unknown

areas of archaeological interest and associated remains. Since the

implementation of satellite imagery in recent archaeological research

proved to be very useful in this context, high resolution satellite tiles

(Quickbird) were used in the study area and subsequently evaluated

regarding their applicability. As demonstrated by the results, the

characteristic Mediterranean environmental conditions (e. g. sparse

vegetation cover, soil erosion, bare rock outcrops) occasionally

impede this prognostic detection to a large extent due to visual

complexity and similar spectral signatures of land surface cover

classes. Thus, we present a data integration approach to identify

ancient settlement locations and useful areas that might reveal new

insights into the colonization history of Crete. Potential archaeo-

logical candidate sites were identified on the basis of a multimethod

strategy using DEMs and Quickbird images. The detected areas are

characterised by favourable environmental conditions, which might

have been used for agricultural purposes, animal husbandry and

settlement activities. Prospective archaeological field surveying can

therefore be conducted time- and costefficiently by immediately

investigating these sites. Even though predictive modelling concepts

such as the one presented have been widely applied in archaeology

for decades, the present study predominantly addresses this topic

from a geoscientific point of view by highlighting a new

geoarchaeological approach for data development, archaeological

remote sensing and GIS analysis.

Taking one step further towards generating a comprehensive and

vivid landscape reconstruction, the Holocene environmental evolution

– particularly during the Bronze-Age – was modeled by terrain

visualisation software packages and geographical information sys-

tems. For this purpose, results from on-site and off-site prospection

in the study area with geophysical and geomorphological methods as

well as digital elevation models provided an appropriate database of

environmental and anthropogenic variables. Within a time series of

several images, the results illustrate the environmental changes that

– according to our geomorphological and archaeological findings –

occurred in Crete during the last 4000 years.

Session 1

73

As shown by the results from Crete, the presented digital tools

perform auspiciously and can be used for a broad range of research

issues, but still their imponderables must be considered. Since they

carry the risk of being manipulative and operator-biased, certain

precautions are required in order to achieve maximum repre-

sentativeness. In addition to the ways of acquiring and processing

appropriate datasets as well as their quality, the amount of input

variables used for the landscape reconstruction is crucial. This

exactly poses a future challenge, because a better understanding of

space and intensified analysis based on the integration of more

environmental data will offer more precise and comprehensive

outcomes. Corresponding tasks explicitly represent the strengths of

geography, which might support archaeological research sub-

stantially. Taking account of these premises, digital geospatial

techniques may significantly contribute to future studies at the

interface between geosciences and humanities. Both used as a single

application in order to consider particular aspects of ancient cultures

and their spatial implications (infrastructures, road networks) and in

combination with further applications (archaeological prospection,

detection of archaeological remains, 3D landscape visualisation) they

help gain better insights into the interactions between man and the

environment. The use of GIS, remote sensing and terrain

visualisation in archaeology thus constitutes the research field of

digital archaeology that offers promising prospects to future in-

vestigations, most notably in Mediterranean regions. Even though

cooperation between archaeology and geosciences is still uncommon

with regard to these methods, recent research points out the

steadily increasing interest in this topic. In addition to geophysical

and cartographical collaboration, computerised prospection is surely

one of the most promising tasks among interdisciplinary geoarch-

aeological research.

Poster

74

SINGATULIN R., YAKOVENKO O.

IT in the reconstruction of ceramics

Research, related to the reconstruction of vessels in its fragments, is

carried out over a long time. Before the advent of computers, all

similar works were reduced to a drawing of the profile of a vessel by

means of displaying its shadow on a screen, and the problem of

recover of the form and the size of a vessel (with the symmetry

property) was solved on one known or on several parameters of the

kept part of the vessel. The situation changed cardinally when

computers appeared. And now, the use of personal computers allows

the involvement of tested mathematical solutions to this problem.

Reconstructions of the forms of receptacles are standard tasks of the

processing of information – dividing the fragments into groups,

search for suitable fragments, sorting, comparison of the textures, of

the colors, etc.

Software and hardware allows the speeding up of the selection of

fragments, their rehabilitation, as well as ensuring the accuracy of

shape and of size of the reconstructed vessels.

The algorithms proposed for solution of these problems are usually

reduced to a full enumeration of pairs of fragments of a certain mea-

sure of conformity the so-called «complementarity». Different

methods are used here: the allocation of the most common sites of

borders of applied fragments, a direct comparison of the visual ob-

jects displayed on the fragments, the method of maximum like-

lihood, etc. These methods are limited a little by its capabilities, be-

cause they cover only one of the aspects of the problem of syn-

thesis. So «advanced» techniques add the global analysis of the

structure to the local analysis of pairs of fragments, which consist of

the connection the fragments to the constructed part. In case of

contradiction the algorithm produces a rollback to the last consistent

state.

However, the serial algorithm, proposed in the «expanded»

technique does not work in the absence of most of the fragments. It

is essential that the algorithms that are included in the above

methods are not able to adapt to the new task. In practice, the part

of the problem material is processed by experts, using a limited set

of heuristics. The systems KBS (knowledge-based systems),

Poster

75

simulating the actions of an expert, are good for solving such

problems. In such a system it is possible to allocate some categories

and to apply them to the problem fragments of ceramics:

analysis of fracture mechanics;

analysis reflecting the properties of the fragments;

analysis of geometric characteristics of boundaries of fragments;

analysis of color, decor, traces, texture of fragments.

Analysis of the destruction mechanics of the vessel allows one to

describe the dynamics, direction of impact force, strain, etc.

Analysis of the reflective properties of fragments of pottery could be

the key when the study is accompanied by a multispectral infor-

mation-measuring system with a wide range of operating frequen-

cies. In an infra-red range, thermal attributes of corresponding types

of fragments of pottery are distinctly allocated.

In the analysis of boundaries, the pair fragments are selected end-

to-end on the same butt length of the sides of polygons

approximating borders. Moreover, the pair fragments must have the

same physical size, such as thickness.

The color characteristics of the fragments provide the largest part of

the attribute information. It is often necessary to produce the

comparisons of fragments only under color histograms.

Optical-geometric characteristics of pottery fragments also give suf-

ficient information, which is defined by the mi rotexture surface, by

the decoration and track entities, of the relative positioning of

fragments. This system architecture of the optical-geometric syn-

thesis is based on the hierarchical streamlining heuristic of the

search of base pairs. Certain categories of base pairs are located on

the top branch of the hierarchical heuristics, which is responsible for

the rough and fast sorting of fragments. On the bottom branch of

the category, additional calculation are demanded for reliability. In

case of false definition, the raising of a fragment upwards on the

hierarchy (rollback) is also monitored and used to adapt by these

algorithms. Such a «mechanism», based on the environment of the

development of the expert systems CLIPS (Common List Processing)

with the connectivity of the package of numerical calculation Mathlab

was realized at the Moscow Technical University. A similar approach

were carried out at the Department of ISiTO (Information Systems

and Technology in teaching), Pedagogical Institute of Saratov State

University in 2006. The program tools of the developed system was

Poster

76

also implemented on the base of the environment of development of

the expert systems CLIPS. An essential difference was the realization

of independent package programs, related to a selective sampling of

fragments of pottery, to the analysis of histograms, to the construc-

tion of three-dimensional models of the vessel, etc. To create the

prototype, the formalization of some heuristics was done. Heuristics

are related to the qualitative analysis of the form of tracelogical con-

stitutions (It is carried out by means of palaeophonographical

technologies).

Fig. 1

A stereophotogrammetric survey of the attribute fragments of cera-

mics was made, on which was based a 3D model of a vessel, filled

with fixed fragments (Fig. 1). Special attention was paid to the deve-

lopment of a custom software interface that allows one in an ac-

cessible way to make automated processing of the fragments using a

flatbed scanner. The results of the produced researches permit us to

conclude that it is expedient to use the package of software

algorithms that simplify office processing of the mass material, raises

the quality and accuracy of processing of archaeological ceramics.

Poster

77

SIOTTO E., VISINTINI D.

3D Texture Modeling of an Important Cycle of Renaissance

Frescoes in Italy

The paper describes the steps of the 3D texture modeling of an

important cycle of Renaissance frescoes located in the Church of

Saint Anthony Abbot in San Daniele del Friuli (North East Italy), from

the surveying by means of a laser scanning and photogrammetric

integrated system to the final photorealistic 3D model in VRML/X3D

computer vision environment.

The first mentioned writings about the church go back to 1308, but

the earthquake of 1348 damaged it seriously and it was therefore

restructured and widened. After these restoration works, concluded

in 1441, the façade was rebuilt with Istria stone in a style resembling

Venetian late Gothic, with a lancet arch portal, a fret-worked rose

window with mosaic stained glass (1470). The church is nicknamed

the “Little Sistine Chapel of Friuli” since it keeps the most beautiful

and harmonious cycle of Renaissance frescoes of the Region, painted

by Pellerino da San Daniele (1467 or 1472–1547). That is why this

pictorial cycle has always raised a great interest and it has been

accurately studied. Giorgio Vasari firstly named it in the artistic

historiography and it has been found in the documents and

chronologies since the XIX century thanks to Fabio di Maniago and

Giovan Battista Cavalcaselle. The pictorial cycle has a complex and

interesting iconographic program including Christological stories, the

four Evangelists, Prophets, Doctors of the Church, stories and Biblical

characters, life episodes of Saint Anthony from Padua and Saint

Anthony Abbot. They are tied up to requirements of the cult of the

brotherhood Confraternita di Sant’Antonio, who has commissioned

the work of art. The frescoes are painted onto the triumphal arch

wall and the adjoining nave walls, the pillars and the intrados of the

presbytery arches, the presbytery and apse walls and domes.

For a detailed analysis of a so important monument, a Terrestrial

Laser Scanner (TLS) system integrated with a photogrammetric

camera surely is the State-of-the-Art surveying technique. Anyway,

once obtained such 3D points, the key problem is the extrapolation

of the geometry of the architectonical elements “imprisoned” within

the cloud with the maximum level of automation, namely the data

Poster

78

“interpretation” making possible the 3D photorealistic modeling. For

the inner and outer surveying of the church, the Riegl Z390I TLS

system integrated with a Nikon D200 photogrammetric digital

camera was employed. The TLS system was placed in suitable four

positions inside the church: from these, fourteen point clouds, with

different axis orientation, angular steps and distance ranges, were

collected in fully automatic way. From the same scanning stations, a

total number of 96 digital images were panoramically acquired with

the photogrammetric camera fixed on the top of the TLS system. In

the same way, the TLS and photogrammetric external surveying of

the principal façade was carried out. Summarizing, 33 millions of

3Dpoints (351 Mb) and 163 photogrammetric images (621 Mb) were

automatically acquired in few hours of surveying! In truth, other

images has been later manually acquired, since some of those

relating the window areas were very dark for exposure troubles;

nevertheless, this has required only one hour of photographic work

with the same camera without the TLS system.

The first step of the data processing, carried out by means of the

RiSCAN PRO® (Riegl) software has concerned the all-together

joining of the various scans (registration) in a unique cloud of 3D-

points.

Afterwards, the points have been processed with suitable tools of

cleaning, filtering, and resampling, and later the surfaces have been

reconstructed as TIN meshes by different commands of

triangulation. Particular care has been devoted to suitably smooth

and decimate such DDSM (Dense Digital Surface Model), in order to

obtain a 3D detailed model but with “few” triangles (tens of

thousands!). In addition, the automatic photogrammetric processes

of cloud colouring, image texturing and ortho-projection or

rectification onto the DDSM have been carried out. This processing

steps are fully automatically carried out, since the camera was fixed

on the TLS top, i.e. it is known the so-called photogrammetric

exterior orientation. This last is instead unknown for the “TLS-less”

acquired images, but it can be easily computed by exploiting a DDSM

textured with the first images and by recognizing, on it and in the

new images, some detail points in the frescoed figures.

The 3D model and the image textures have been exported as a WRL

file, so allowing interactive tours in the VRML/X3D virtual reality

space. The obtained VRML/X3D model is a computer science

Poster

79

instrument useful both to experts and to real or virtual visitor, since

the model will be accessible by web. It will be possible the free 3D

exploration and/or the choice of various thematic tours (chronologic,

restoration, technique, geographical and in the time), with a virtual

camera going exactly in front of the frescoed scene and subjects

(without angular roll) and at a variable distance in order to wholly

observe and them.

Furthermore, many VRML/X3D anchors have been suitably created,

so to link each frescoes scene with the corresponding informative

card of the web databases of S.I.R.Pa.C. (Regional Information Sys-

tem of the Cultural Heritage – http://www.sirpac-fvg.org/index.asp)

or Ar.I.S.T.O.S. (Database for the History of the Historical-Artistic

Objects Preservation – http://aristos2.mbigroup.it/) or S.I.Ca.R.

(Information System for the Restoration Yards – http://sicar.

mbigroup.it). This work of 3D modeling is thought in fact for a web

main use and fits within the regional project Computer Sciences and

Web for the Cultural Heritage: Mobile and 3D innovative services for

the Tourism of the University of Udine (http://www.infobc.uniud.it).

This project provides the creation of a complex GIS collecting and

integrating existing and new numerical 3D models and thematic

database, for the management, protection, utilization, and

enjoyment of the historical-artistic and archaeological heritage of the

Friuli Venezia Giulia Region. Screenshot of the 3D textured model of

the frescoes of St. Anthony Ab. Church in VRML/X3D environment.

Poster

80

SLIZEWSKI A.

NESPOS – Digitalizing Pleistocene People and Places

NESPOS (www.nespos.org) is an online database for archaeologists

and anthropologists that opens new possibilities of data sharing and

communication. The supporting organisation is the NESPOS Society

e.V. and it is located at the Neanderthal Museum in Mettmann,

Germany.

NESPOS was the result of the 24-months-project “TNT – The

Neanderthal Tools” which started in 2004 and was founded by the

European Union. The project was initiated by the Neanderthal

Museum and accomplished in close cooperation with the University

of Poitiers (France), the Croatian Museum of Natural History

(Zagreb), the Royal Belgium Institute of Natural Sciences (Brussels)

and the technical partners Hasso Plattner Institute, PXP and National

Geographic (Semal et al. 2004, Gröning et al. 2005).

NESPOS is an open platform, designed for paleoanthropologists and

archaeologists worldwide who can up- and download research

results. All data can be protected in a membership area. More

sensible data can be stored in personal security spaces which are

only accessible by the members who created them. For international

working groups, a security space can be made accessible to several

researchers. An example of such a working platform within NESPOS

is the European Virtual Anthropology Network (www.evan.at).

Within NESPOS all related data is linked to each other. By calling up

e.g. a single fossil you also get information on all other human re-

mains found in the same place, PDFs of literature, site data such as

stratigraphy, research history, fauna, archaeological features and

artefacts.

The software package VisiCore allows creation and handling of

6faces (Berens, Slizewski 2008) as well as voxel, polygonal, single

picture models and landscape or site models.

At the moment, the NESPOS membership area includes photos,

literature, CT data, surface scans (see e.g. Slizewski, Semal 2009)

and 6faces of human fossils and artefacts from more than 70 Asian

and European sites. Within an ongoing research project 50 sites from

the Iberian Peninsula will be added within the next two months.

Poster

81

NESPOS was originally designed for a limited number of researchers

and Neanderthal fossils. But during the last five years, the database

increased so much and proved to be such a good research tool that

in February 2009 an updating process was launched. Main goal was

a better usability of nespos.org to attract scientific users to provide

more data and to satisfy the growing public interest in NESPOS. On

19th

June 2009 the new version of NESPOS went online. In the

course of this relaunch, the database was broadened to all Pleisto-

cene sites and human fossils, now also including Palaeolithic art (Breuck-

mann et al. 2009), Australopithecines, Homo erectus, anatomically

modern humans and a huge collection of modern reference data.

Basic information on sites and fossils are now available to the public

in form of a digital encyclopaedia on the Palaeolithic. Only high

resolution 3D data and personal or security spaces remain restricted

to scientific members.

Embedded links to Wikipedia, interactive GoogleEarth maps, a

science daily feed and the possibility to create a RSS feed for

NESPOS complete the new Web 2.0 functionality of NESPOS.

Poster

82

SOTIROVA K.

Multimedia Galleries of Cultural Heritage – Piece of a Puzzle

or the Overall Picture?

2009 has been declared by EU as European year of creativity and

innovation. 2008 was year of the intercultural dialogue. The current

paper is trying to explore these European priorities in the context of

Digital Humanities research field with stress on various interactive

multimedia presentations of cultural heritage, EduTainment (techno-

logies).

The research topics in this paper is how we (scientific groups in

Digital Humanities area and cultural heritage holders) are addressing

intercultural content in our common, as we say, European heritage,

what story we offer by the way this content is presented online and

how multimedia EduTainment tools might challenge us and show

an innovative way for richer e-presentation.

Among the questions we are asking is what knowledge the

knowledge society of 21c. represents online and whether technology

development is what the cultural heritage holders and science

centers are missing to get full attendance. The technology (2D and

3D reconstruction, Computer Vision applications, Computer Tomo-

graphy, excavation's historical documentation from multimedia data

etc.) in general is developing so quickly that cultural heritage holders

are lacking time to make the digital recreation of their collections

more human-like and closer to the real visitors. Thanks to techno-

logy the digital heritage from Antiquity till today is more vivid, but

the storytelling, which is the most important aspect of all computer

generated worlds needs to be further developed. What needs to be

such a development? In our opinion it needs not more attractive-

ness, but more depth; depth in terms of challenging questioning and

different points of view presented, incl. the point of view of the

different Other (in historical context this is the Neighbour, the

Enemy, the political Partner, the opponent etc.).

Cultural heritage is one of the best reservoirs for different

multimedia EduTainment tools (playful learning) and games. The

virtual environments (especially cultural heritage ones) can learn a

lot from game design. A good example of dialogue between

technology and humanities is Virtual Warrane application and

Poster

83

Haemimont Ltd. European history based games. They both illustrate

a broader tendency – to capture indigenous knowledge on computer

systems, which can then be used to protect, preserve and promote a

specific-history-based heritage (culture). Protection, preservation

and promotion of cultural heritage is the first step of what we call

interactive intercultural dialogue with the different Other. The next

step is attractive e-presentation.

Story-telling is the most important element here and in any

playfull-learning application. The success of the best sold ever com-

puter game, SIMS, is interesting example, since it is based on a

story, which simulates the real life, where the player is not God-like

personage (like in many other games), but in some sense he is

limited by the actions of his ‘heroes’. Second life is another example,

where the search for ‘better-life-simulation’, which has to be closer

(and different at the same time) to the reality is very obvious. Con-

cluding, we assume that if the story told offers the player/museum

visitor to acquire some skill or/and knowledge while being in the

‘magic circle’ (K. Zimmerman) of the gameplay, then the goal is

achieved.

WEB2.0 and 3.0 and esp. domain ontologies are offering the techno-

logy for semantic search trying to satisfy the hunger for classified

information and tangible net. For now we cannot offer case studies

where semantic search is already incorporated in cultural heritage

related web site. The status quo shows that museum collections,

presented in institutional web portals in different countries show

one-and-only viewpoint *policy*. ‘Good’ example is British museum,

big part of whose collections are set out of their original contexts

(original land, culture, language). Another ex. is EUROPEANA portal,

where ‘view in original context’ link shows (in most of the cases) a

picture, i.e. the digital image without any text. The original context,

historical and cultural, is always presented inseparably from the

point of view of the presentation-author, i.e. object-owner. Inter-

cultural and interdisciplinary dialogue as well as knowledge

objectivity require showing at least one more *picture* of the same

heritage object.

Usually opposing multimedia presentations are absent in the web

portals of European museums. This gap can be perfectly filled up by

EduTainment multimedia tools. Serious storytelling, which we define

as interdisciplinary knowledge based (game)play is an alternative

Poster

84

and educative way for making a virtual gallery of linked stories, not

only mechanical sum of well textually explained heritage objects.

Imagine BAUHAUS (1919–1933), the most important school of

architecture, design and art

in the twentieth century,

presented in Weimar

Museum ONLY by its Fine

arts? Although showing the

context in Fine arts domain

is relatively easier from

technology and curators

point of view (see PRADO

PLAY section). It is much

more complex when the

heritage object is a relic from a war. Perfect bad example of one-

and-only point of view is Cold War Museum and Imperial War

Museum London. Their collections are relevant to various cultures,

but do not present intercultural links, nor opposing storytelling.

The PhD thesis of the author is following in practice the sketched

above scenario for presenting heritage along with challenging story-

telling, following ‘the rules of play’; the multimedia and Semantic

web technologies here are necessity, but the crucial part is Homo

Sapiens Ludens. Whereas in games world there is statistics for the

different profiles of users, such statistics is missing for cultural

heritage holders web sites. This is the first point we have to start

from in order to create more successful museum applications. As a

conclusion we would say that European (and not only) web portals,

combining collections of different museums, galleries and libraries

are hoping that Semantic web technologies will ‘do the work’ of

linking logics and semantics of the objects presented, as the story

telling in game applications do. We think though that such a hope is

over above the possible one technology can do, and we do need the

human storytelling, combining (in a playful way) different points of

view. Such combinations nowadays are missing and not thought

enough, even in research labs.

Session 6

85

UHLIR Ch., UNTERWURZACHER M., SCHALLER K.

Historic Quarries – the Database and Case Studies

Introduction

Historic quarries (HQ) as material sources for monuments,

architecture and consumer goods are part of our archaeological and

industrial heritage as well as cultural landscape. HQ´s are in danger

due to many factors: garbage dumps, modern quarries, enlargement

of urban areas, vandalism, looting, etc. At the moment its heritage

value is not properly recognized nor protection concepts are

developed. Recent investigations on antique quarries of Egypt show

that about 1/3 already have been destroyed within the last three

decades.

Historic Quarries, a European project led by the CHC – Research

Group for Archaeometry and Cultural Heritage Computing of

Salzburg University/Austria, is being implemented to collect sample

data and build up a database on a large number of individual HQ

sites and related monuments in Central Europe. The data comprise

historical and technical information, site and stone related data

(petrography) completed by images of the sites (historical views and

current use of the sites) and information about the historic

destination of the mined material (historic monuments, distribution

in Europe).

Definition

A historic quarry is a defined mining area within a suitable resource

of natural stone containing remains of the different mining processes

such as tool marks, dumps, semi-finished goods, infrastructure,

remains of workshops, tools as well as accommodations and social

facilities.

The Database

The petrographical, geotechnical and geochemical data of quarries

used in history as reference groups for provenancing monuments are

mostly unpublished or hidden in “grey literature”. These data which

also include photographs and maps should be made accessible to the

entire research community by the interdisciplinary information

system www.saxa-loquuntur.org.

Session 6

86

It consists of two main databases:

The quarry database: describes general quarry information,

localization, material, geological information, dating of quarrying

phases, quarry morphology, signs of treatment, historic infra-

structure, semi finished goods, archaeological findings, authors and

literature The sample database for quarries and monuments: de-

scribes sample information, material, macroscopical -, microscopical

-, geochemical -, X-ray data, material technical properties, authors

and literature. Because of the flexible structure of the analytical

section new methods easily can included.

For the communication between archaeological sciences and natural

sciences a simplified interactive rock thesaurus was developed on

the base of the IUGS rock nomenclature. Controlled vocabulary,

editable by content administrators, for various allows was

established.

Interlink between the sample database to various monument

databases enable a full interdisciplinary monument analysis.

At final stage the databases can be queried by simple and advanced

search methods. The information system will provide visualisation

tools for geochemical data, a photo board for the comparison of thin

sections and a cartographical visualisation of the search results.

Currently the quarry and sample database contains mainly

information and data of Roman used marbles from the Alpine and

Carpathian region. Within the course of the project the area of the

former Austrian Hungarian Empire will be examined.

Core Data

As result of a first data exploration within the project area an

amount of about 10.000 sites have been identified. As sources

mainly databases of national surveys and historic material collections

were used. To manage that huge amount of data within the time

frame of the project a dataset called “core data” has been

developed. This data set involves the name of the quarry or quarry

district, the physical localization by coordinates and the hierarchical

system of the country (Loc.Name/village/county/county), general

material information, a rough chronology of the quarry activities (if

known) and related literature. These core data sets will be the base

for further data exploration within follow-up projects.

Session 6

87

Outstanding Quarries

For each country full datasets will be collected for “outstanding”

quarries. The identification and selection of outstanding quarries will

be done by evaluating their historical significance using a system

which was developed for physical cultural heritage:

• Associative / symbolic value: a quarry itself and cultural

remains found in a quarry can deliver important cultural

information on the past and can be connected with the

collective memory of the people of adjacent areas.

• Informational value: using a multidisciplinary approach

various experts of different scientific fields investigate

quarries showing the “status quo“ on resources and the

overall locality, providing results that are suitable for further

investigations.

• Aesthetical value: the aesthetical value of a quarry can be

seen in combination with natural and manmade influences

on a resource in respect of its present-day appearance and

its development over time.

• Economic value: from the economic point of view often cost-

benefit analysis are made. Thus decisions on cultural

resources concerning conservation, research, exhibitions,

decay and destruction of a quarry landscape also have an

economic dimension.

• Social and spiritual value: this value is related to reverence

for the place. It is connected with room use for social

events, for pressure groups and associations.

For the selection of outstanding quarries also quarry specific para-

meters like material, time and space dimensions and the connection

with outstanding monuments will be used.

Case Studies

For the first case studies within Austria the quarry districts of the

Adnet - and Untersberg Marble have been chosen.

Session 4

88

VAR P., PHAL D., NGUONPHAN P., WINCKLER M. J.

3D Reconstruction of Banteay Chhmar Temple for Google

Earth

The Banteay Chhmar temple is one of the most significant Hindu-

Buddhist temples in Cambodia, established during the reign of King

Jayavarman VII in the second half of the 12th century and dedicated

to the king's son Srindrakumara (Sanday 2007). In the course of

time almost all of this magnificent temple has been ruined, making it

one of the most mysterious of the Khmer temples. Through the work

of the Global Heritage Fund, preservation and reconstruction of the

temple area were recently started. But people only begin to under-

stand (since the start of the project in 2007) the structure, the

concept and the amazing ancient Khmer architecture represented in

this building complex. A full 3D computer model, as proposed in our

work, for a temple as vast and diverse as this, is a great opportunity

to fill this gap. In cooperation with Google Earth we try to recreate a

model for the understanding of the whole area complex, featuring

the full scale of details of the original architecture and the stone

materials used, in order to explore structural and artistic properties

of the Banteay Chhmar style.

The basis of the project is formed by the work of Nguonphan. The

surveys of structural parts (see fig 1, 2) were done in high geometric

detail using the Angkor Temple Generator (APG) (Nguonphan 2009).

However, this model is geometrically too complex for a represen-

tation in online interactive viewing environments due to the immense

amount detail. The currently available models of similar buildings

(e.g. Google Earth) on the other hand feature only very little of the

geometry (e.g. Angkor Wat, see fig. 3; Krzysio 2007) and lack

essential information about Khmer architecture and style. In our

project we therefor will generate a virtual reconstruction of Banteay

Chhmar which is acceptable in geometric detail, architectural quality

and technical feasibility.

Using our experience from work at the IT center at the Royal

University of Phnom Penh (RUPP, Cambodia) and the Interdis-

ciplinary Center for Scientific Computing (IWR) at Heidelberg

University (Germany), we will incorporate methods from the Angkor

Temple Generator (based on AutoCAD), Google SketchUp and

Session 4

89

Google Earth to accomplish this task. The support from the Google

Workshop GCAMP@RUPP in June 2009 is also a major source for the

completion of our goals.

Figure 1–3

Our overall purpose of creating a 3D model of Banteay Chhmar is to

build a bridge between specialized techniques of information

technology and an application in the humanities, leading to a pro-

duct to educate people all around the world about ancient Khmer

architecture and the power of the ancestors of the Kingdom of

Wonder.

Poster

90

WAND E.

Hidden in the Sand! – The Famagusta Research Project.

One Land, One City, One Memory

The goal of the Famagusta Research Project is to call attention to a

period when the history of divided Cyprus stood still. At the same

time, however, it also intends to have an influence – directly or

indirectly – on people and protagonists: those who bear the respon-

sibility for the current state of affairs, those who may possibly have

had to experience this tragedy at first hand and those who hitherto

have been no more than mute spectators and remain mere extras –

in order to change their way of thinking, knowing and acting.

Thesis

Can a multimedia research project influence minds and effect

political changes? Can it touch hearts and effect emotional changes?

Will it be possible that latest communication technology enables the

former residents of Famagusta – now a refugee community scat-

tered world-wide – to get linked together in an online community

which maintains its demand after reunification through a virtual

environment.

Session 4

91

WULFF R., SEDLAZECK A., KOCH R.

3D Reconstruction of Archaeological Trenches from

Photographs

Summary

In this work, an image-based algorithm for the 3D reconstruction of

archaeological trenches is proposed. We extend the structure-from-

motion approach for general rigid scenes described by Pollefeys et

al. (2004) to better fit the archaeological needs.

The algorithm requires a calibrated digital camera for image ac-

quisition and a set of 3D points with well-known world coordinates.

The 3D points are needed to transform the scene into the absolute

coordinate system that was used at the excavation site. This allows

for measurements in the 3D model or for fusion with other objects

that share the same coordinate frame. In addition, a new algorithm

to minimise errors in pose estimation is introduced that exploits prior

knowledge about camera movement.

Background

When working in archaeological excavations, a detailed documen-

tation of the finds and features is very important. The main reason

being that the configuration will be destroyed when unveiling the

next layer. A wide variety of techniques is used for documentation,

including drawings, photography and photogrammetry, most of them

being very time-consuming. Although in photogrammetry 3D data is

already used, none of these methods aims at producing a complete

3D model of the trench. However, a 3D model can be very helpful in

the interpretation of the finds and features, because a 3D repre-

sentation is intuitive and pleasant to human perception. Further-

more, a 3D model allows for metric measurements in 3D space, even

after the destruction of the configuration, and can be used for

presentations to the general or academic public, e. g. in museums.

Recently, some methods have been proposed to reconstruct 3D

models in the field of archaeology. The designated use of the

methods introduced by Tsioukas et al. (2004) and Zheng et al.

(2008) is the reconstruction of small finds, so they are not applicable

to reconstructing trenches. A method for large-scale reconstruction

was suggested by Ioannides and Wehr (2002). The approach is

Session 4

92

based on laser scanners, and detailed models can be achieved.

There are two main drawbacks of this method: First, expensive

equipment is needed which is not part of the general documentation

process in archaeology. Second, the data acquisition can be very

time-consuming: The survey of a cave of size 7.2 x 7.2 x 4.35 m

took about 1.5 hours.

Our Work

The proposed method extends the general structure-from-motion

approach described by Pollefeys et al. (2004) to better fit the

archaeological needs, similar to the approaches taken by the 3D

Murale project (http://dea.brunel.ac.uk/project/murale/) and by the

ACVA'03 workshop (http://www.lems.brown.edu/vision/conferences/

ACVA03/). Based on an ordered sequence of photographs with

overlapping viewports in successive images, a 3D model represented

by a triangle mesh is generated. As a first and crucial step, the

special archaeological needs had to be figured out. This was done in

close collaboration with archaeologists of the University of Kiel,

Germany.

The main requirement was to allow for measurements in the recon-

structed model. In the classic structure-from-motion approach the

scene is reconstructed in an arbitrary coordinate system, so that its

scale, position and orientation can hardly be predicted. To allow for

measurements, we transform the scene into the absolute coordinate

system that was used at the excavation site. Additionally, the model

can easily be fused with arbitrary 3D objects that share the same co-

ordinate frame, e. g. with models of other trenches from the same

site. To compute the transformation, a set of 3D points with well-

known coordinates, so called photogrammetry points, are required.

These points are already part of the documentation process, e. g. for

computing pseudo-orthographic photographs, and can be reused

here.

Exploitation of prior knowledge about camera movement leads to a

new algorithm to minimise errors in pose estimation. If follows a

simple approach and we call it LoopClosing. The exploited assump-

tion is that the camera was moved in an orbit around the trench,

pointing inwards. Furthermore, we expect the first and the last

image of the sequence to overlap in the same range as all other

images of the sequence. This enables us to append the first image

Session 4

93

again at the end of the sequence. Then the geometry between the

last and the appended image can be estimated. In theory, the poses

of the first and the appended image will be identical, but the

estimation will yield a discrepancy between them. This discrepancy is

distributed amongst the poses of all images according to a weighting

function.

Figure 1: Visualisation of the reconstruction process. a) An input image. b) The

estimated camera poses and a sparse point cloud of 3D keypoints. Each pyramid

represents a camera. c) A depth map (dark=near, light=far, black=undefined). d) The

final model fused from four views.

The reconstruction algorithm iterates over the whole image se-

quence. In each step, distinctive keypoints are detected auto-

matically in the current image. These keypoints are then compared

to the two predecessors of the current image to obtain a set of 2D

correspondences. To initialise the structure, the 2D correspondences

of the first two images are used to estimate the epipolar geometry

between them. Based on that, a sparse set of 3D points is estimated.

These points are needed to compute the poses of the remaining

images using 2D-3D correspondences. After all input images have

been processed, the LoopClosing algorithm is applied. To minimise

the reprojection errors, a global bundle adjustment is performed

afterwards. Now we have a complete reconstruction of the camera

Session 4

94

path. The next step is to transform the scene into the absolute

coordinate frame mentioned above. Then for each view a dense

depth map is obtained by applying a multi view stereo algorithm

between each two successive views. From each depth map a 3D

model is built, textured by the corresponding input image. The last

step is to fuse these models to compensate for occlusion.

Outlook

The long term objective of the project is to develop a complete

system for 3D re-construction with special consideration of the needs

in archaeology, offering a tool for 3D measurement, segmentation,

and detailed visualisation of archaeological excavations.

Session 1

95

YARLAGADDA P., MONROY J. A., CARQUÉ B., OMMER B.

Towards a Computer-based Understanding of Medieval

Images

The grand goal of computer vision is to enable computers to auto-

matically understand images so that they can reason about objects

rather than being limited to a mere analysis of individual pixels.

Gaining access to the semantics of images is a necessary pre-

requisite for computers to support human users in many complex

tasks such as object search in image databases. In contrast to this,

current object retrieval systems that are bound to a text-based

search are significantly limited by the ambiguity of textual an-

notations. Problems of completeness and compatibility arise from dif-

ferent taxonomies, which are a result of different scientific cultures

and of the diverse goals that different observers of an image have.

Current algorithms for object retrieval require textual annotations of

images which fundamentally limits their usability due to a number of

reasons: i) A user can only find entities that have been previously

deemed important by the annotator of the dataset. ii) Textual

annotations do not provide localization information so that object

detection and reasoning about the (spatial) relationships between

objects becomes impossible. iii) Annotating the dataset is labor

intensive and so the approach is limited to small sets of images. To

deal with these shortcomings we present a system that can find

objects and analyze their variability directly in the images. In

contrast to text-based search, our approach learns object models

from few training images. Thereafter, the method can automatically

detect new object instances in large collections of novel, unlabeled

images.

Our contribution in this work is threefold, concerning i) bench-

marking, ii) object analysis, and iii) recognition.

Benchmarking: We have assembled a novel image dataset with

groundtruth segmentations that is highly significant for the

humanities due to its unusual completeness of late medieval

workshop production as well as for computer vision since it is the

first of its kind to enable benchmarking of object retrieval in pre-

modern tinted drawings. The primary source of our research are 27

late medieval paper manuscripts from Upper German origin archived

Session 1

96

by Heidelberg University library. These codices are illustrated with

more than 2,000 half- or full-page tinted drawings. We start from

object categories which are represented in sufficient numbers and

which have, at the same time, a high semantic validity since they

belong to the realm of medieval symbols of power. That way, we can

ensure right from the start that our approach has the highest

possible connectivity to the research questions of those disciplines

which work on medieval images – notably art history and history

with a focus on ritual practices or on material culture. To establish

groundtruth for object localization and classification, a tool that

gathers manual segmentations for datasets of medieval images has

been developed and applied to the dataset.

Object Analysis: Images are decomposed into preparatory drawings

(see illustration) and the coloration. That way, boundary information,

which is suitable for a shape-based representation, is separated from

appearance, which can be degraded significantly due to the aging of

images. To represent objects, the shape of extracted boundary

contours is described by capturing their local orientation distributions

in histograms on multiple coarseness scales. This representation is

robust to local distortions in the image and due to its multi-scale

nature, it captures characteristic, small details of objects. Based on

this object representation, our approach allows to analyze the

variability and interrelatedness of different instances of a category

and it gives scientists from the humanities a visualization of the

structure of object categories. In a first step, hierarchical, agglo-

merative clustering using Ward’s method (minimum variance

clustering) filters out duplicate samples from an object category.

Projecting the shape representations of these samples into a 2-D

subspace using multidimensional scaling (MDS) provides a direct

analysis of the relations between all the instances of a category in

the database in a single visualization (see illustration). Such an

automatic analysis and its visualization is of essential value for the

humanities as it combines data from large numbers of images that

could not be compared by a researcher. From the standpoint of the

humanities, the projection successfully clusters category instances

according to their semantic structure and relatedness and it brings

order into the high variability within a category. In particular, our

illustration for the category “crown” shows that to the simple crown

circlet further elements like arches, hats, or helmets are added.

Session 1

97

Current text-based classification systems do not take into account

this diversity of types, so that only the keyword ”crown” is

searchable. Since automated image-based search does not suffer

from the polysemy in annotation taxonomies, it becomes a crucial

instrument to assist with the detailed differentiation of subtypes. By

clustering segmented crowns according to the criterion of formal

congruencies, distinct clusters become apparent. These groups

feature different principles of artistic design, which are characteristic

for different teams of artists that have drawn the images. Group (A)

indicates the concise and accurate style of the Hagenan workshop of

Diebold Lauber; group (B) the more delicate and sketchy style of the

Swabian workshop of Ludwig Henfflin. The detection of specific

drawing styles of artists is a highly relevant starting point to

differentiate large-scale datasets by workshops, single teams within

a workshop, or even by individual draftsmen.

Session 1

98

Recognition: Based on our shape representation, we propose a

category-level retrieval system that can detect and classify objects in

novel images. Therefore, the shape descriptors in a test image are

compared against those from training images. Each descriptor then

casts a vote for the object location and category in the test image to

perform object detection. Our recognition results (see illustration)

show the potential of the approach.

The present case study on the Upper German illustrated manuscripts

of Heidelberg University library exemplarily demonstrates the deep

insight into medieval object representations and their artistic context

provided by automated image analysis: in their extensiveness and

detailedness, these observations are unmatched by the common

methods which are, up to now, available to research on cultural

heritage.

Poster

99

YULE P.

Laser Scanning of a Mauryan Column Complex in

Sisupalgarh, an Early Historic Fortress in Coastal

Orissa/India

Renowned in the context of Ashokan India (4th century BCE),

Sisupalgarh, the largest early historic fortress in the eastern part of

the Subcontinent (with exception of mere traces of Pataliputra,

present-day Patna), plays a role in virtually all discussions bout this

period. Its symmetrical plan and great size (130 ha, 1190 x 1150 m,

measured at the top of the glacis) reveal an architectural ideal for its

day. South Asia experts usually discuss it as an example of defensive

early history architecture, largely omitting any relation to prede-

cessors, relatives, or successors. Recent research conducted by a

team from the University of Heidelberg, Utkal University in

Bhubaneshwar and the University of Applied Sciences in Mainz has

rekindled the research largely of the 1940s, revealing the uniqueness

of Sisupalgarh and its role in the eastern part of India. To our

knowledge this is the first application of this kind of scanning in the

archaeology of the Subcontinent.

A mysterious group of columns and a hill, considered to be part of a

palace, were hardly even photographed although they were known

since the early 20th century. In 2005 we recorded the site shoal

khamba (16 columns) by means of a tachymeter and a laser

scanner.

Poster

100

The recording of the column complex and topography rests on some

15 million measurements recorded at a rate of nearly 1000 points

per second–a veritable cloud of laser-measured points. In order to

achieve a complete coverage, 20 scans were taken from different

observation stations. The single scans were registered to form a

common point cloud and thinned to the necessary resolution,

resulting in 3 million points which were connected by 6 million small

triangles which actually describe the surface. A dot-reducing

algorithm made the calculation manageable. A final step was to

render and animate the recorded data in order to make it “come to

life” and to convey render its spatial appearance more vividly than

possible simply with elevational lines.

The complex can be shown in any of a variety of ways including

animation.

Session 2

101

ZAMBANINI S., HERRMANN M., KAMPEL M.

An Automatic Method to Determine the Diameter of

Historical Coins in Images

Numismatics deals with various historical aspects of the

phenomenon money. The basic objects that numismatists deal with

are historical coins and in the numismatic community coins are

identified through their description, i.e. a set of features specifying

the type, minting place, etc. of the coin. Usually a coin description

also includes physical values like weight and maximum diameter.

This paper presents an automatic method for computing the

maximum diameter of coins by means of a ruler placed next to the

coin. The method consists of two steps:

1. Determination of the spatial resolution of the image, i. e. the

pixels’ real world size, from the ruler visible in the image.

2. Segmentation of the coin and computation of the maximum

diameter.

Such a method is of high interest for daily practical work of numis-

matists since it allows a much faster processing of coins. Ad-

ditionally, an automatic image-based measurement is more accurate

than a manual one using a caliper. Another need satisfied by an

automatic determination of the spatial resolution is the fast

preparation of coin images for publication. Usually coin images are

required to be published in a 1:1 scale to their real world size and

with a white background. The proposed method is able to deal with

an arbitrary orientation of the ruler and placement of the coin. Thus,

only minimal constraints on the image acquisition are necessary.

Usually, a camera mounted on a camera stand providing a constant

distance and parallelism between the coin and the camera’s image

plane is sufficient.

For the computation of the spatial resolution the Fourier transform is

used which is motivated by the fact that the ruler marks form a re-

gular pattern that produces ridges in the frequency domain of the

image. Ridges are detected in the power spectrum corresponding to

the marks on the ruler. The information about the ridge location in

the power spectrum allows calculating the frequency of the ruler

marks and thus the spatial resolution of the image. As a final step,

the inverse Fourier transform is applied solely to the ridges found

Session 2

102

before. This allows to determine the location of the ruler in the pic-

ture which can be used to give visual feedback to the user later on.

Coin segmentation is based on the assumption that the coin itself

possesses more local information content and details than the rest of

the image, i. e. the background. Therefore two filters are applied on

the image to highlight regions with high information content: the

local entropy and the local range of gray values. Local entropy

derives the measure of local information content from local gray

value histograms, whereas the local range of gray values is defined

as the difference of the maximum and minimum gray value of a local

neighborhood. The outputs of these two filters are summed-up to

build the final intensity image where a thresholding is applied on.

The method was evaluated on a set of 30 images gathered from the

Romanian National History Museum. For each image the ground

truth was obtained by manually segmenting the coin and measuring

the distance between ruler marks using a commercial image editing

program. Since the final output is the maximum diameter of the coin

this is determined by computing the maximum distance between

border points of the manual and automatic segmentation.

With an average error of 0.49 % for the segmentation and 1.00 %

for the spatial resolution determination it can be concluded that both

methods give accurate results for the given dataset. This also leads

to an accurate measurement of the real coin diameter with an

average error of 0.20 mm or 1.19 %. Furthermore, a maximum error

of 6.64 % emphasizes the robustness of the proposed method. The

robustness of the spatial resolution determination is also indicated by

a maximum error of 6.67 %. However, on the given dataset all rulers

were placed vertically in the image. Therefore, to verify the

robustness of the method for arbitrary rotated rulers, 10 randomly

selected images were rotated in 20 degree steps. Thereby, we

obtain 18 individual measurements for each image and the error is

evaluated by means of the coefficient of variation, i. e. the standard

deviation of the samples divided by their mean. As a result of the

experiment, the average coefficient of variation for all 10 images lies

at 0.52 % which shows the low sensitivity of the method to the ruler

orientation.

In summary, experiments have shown the gain of the method for

numismatists by improving both the accuracy and speed of coin

processing. The method is fully automatic except for the one-time

Session 2

103

manual adjustment of the unit length between two ruler marks for a

given ruler type. Since the proposed method for determining the

spatial resolution of images is in theory applicable on any type of

ruler, the method can be used to measure or scale other kinds of

objects as well, like skin lesions or initials and letters of ancient

documents. Though the experiments proofed the general robustness

of the method, in the future a more comprehensive evaluation with

images from several sources is planned to underline the usefulness

of the method for the numismatic community.

Notes

104

Notes

105

Notes

106

PARTICIPANTS

Participants

108

PARTICIPANTS

page

ALTENBACH Holger

Institut für Ur- und Frühgeschichte

und Vorderasiatische Archäologie

Ruprecht-Karls-Universität

Heidelberg

46

ARNOLD Matthias

Historisches Seminar, Zentrum für

Europäische Geschichts- und

Kulturwissenschaften. Ruprecht-

Karls-Universität Heidelberg

67

BAATZ Wolfgang

Institute for Conservation and

Restauration

Academy of Fine Arts Vienna

6

BALZANI Marcello

DIAPReM – Department of

Architecture, University of Ferrara

8

10

BATHOW Christiane

Breuckmann GmbH, Meersburg

62

BERNER Konrad

Historisches Seminar, Zentrum für

Europäische Geschichts- und

Kulturwissenschaften. Ruprecht-

Karls-Universität Heidelberg

67

BOCK Hans Georg

IWR – Interdisciplinary Center for

Scientific Computing, University

Heidelberg

BOOS Silke

Institute for Spatial Information

and Surveying Technology, Uni-

versity of Applied Sciences, Mainz

12

BOU Vannaren

Information Technology Center

Royal University of Phnom Penh

45

BREUCKMANN Bernd

Breuckmann GmbH, Meersburg

62

page

CARQUÉ Bernd

IWR & Transcultural Studies

University of Heidelberg

95

CARROZZINO Marcello

IMT – Institute for Advanced

Studies Lucca

56

CASSELMANN Carsten

Institut für Ur- und Frühgeschichte

und Vorderasiatische Archäologie

Ruprecht-Karls-Universität

Heidelberg

46

CHRIST Georg

Transcultural Studies

University of Heidelberg

17

EITEL Bernhard

Geographical Institute – Labora-

tory for Geomorphology and Geo-

ecology, University of Heidelberg

71

FELICETTI Achille

PIN, University of Florence, Prato

53

FELICIATI Pierluigi

MAG Committee

University of Macerata (IT)

20

FEO Rosalba De

Soprintendenza per i Beni Archi-

tettonici e per il Paesaggio, il Patri-

monio Storico, Artistico e Demo-

etnoantropologico per le Province

di Salerno e Avellino, Salerno

8

FERSCHIN Peter

IEMAR – Institute of Architectural

Science/Department Digital

Architecture and Planning

Vienna University of Technology

22

FLÜGEL Christof

Landesstelle für die nicht-

staatlichen Museen in Bayern

Munich

24

Participants

109

page

FORNASIER Massimo

RICAM, Austrian Academy of

Sciences, Linz

6

GALVANI Guido

DIAPReM – Department of

Architecture, University of Ferrara

10

GIETZ Peter

Historisches Seminar, Zentrum für

Europäische Geschichts- und

Kulturwissenschaften, Ruprecht-

Karls-Universität Heidelberg

67

GOLDMAN Thalia

Department of Physics of Complex

Systems, The Weizmann Institute

Rehovot

26

GORDAN Mihaela

Faculty of Electronics, Telecom-

munications and Information

Technology, Technical University

of Cluj-Napoca

49

GROSMAN Leore

Department of Physics of Complex

Systems, The Weizmann Institute

Rehovot

26

HAUCK Oliver

Fachbereich Architektur

Technische Universität Darmstadt

28

HEIN Anno

Institute of Materials Science,

N.C.S.R. “Demokritos”, Aghia

Paraskevi

30

HERMON Sorin

Science and Technology in

Archaeology Research Center, The

Cyprus Institute, Nicosia

53

HERRMANN Michael

Institute of Computer Aided

Automation

Vienna University of Technology

101

page

HOHMANN Georg

Referat für Museums- und

Kulturinformatik, Germanisches

Nationalmuseum, Nürnberg

32

HOPPE Christoph

IWR – Interdisciplinary Center for

Scientific Computing, University

Heidelberg

35

HÖRR Christian

Fakultät für Informatik

Technische Universität Chemnitz

37

JÄGER Willi

IWR – Interdisciplinary Center for

Scientific Computing, University

Heidelberg

JUNGBLUT Daniel

Goethe-Center for Scientific

Computing, Goethe University,

Frankfurt am Main

41

KALASEK Robert

Department of Spatial Develop-

ment and Infrastructure &

Environmental Planning, Vienna

University of Technology

62

KAMPEL Martin

Institute of Computer Aided

Automation, Vienna University of

Technology

101

KARL Stephan

Landesmuseum Joanneum, Ab-

teilung Archäologie & Münz-

kabinett, Schloss Eggenberg, Graz

41

KILIKOGLOU Vassilis

Institute of Materials Science,

N.C.S.R. “Demokritos”, Aghia

Paraskevi

30

KLEPO Višnja

Institut für Architektur- und

Kunstgeschichte, Bauforschung

und Denkmalpflege

Vienna University of Technology

43

Participants

110

page

KOCH Reinhard

Institute of Computer Science

Christian Albrechts University, Kiel

91

KOR Sokchea

Institute of Computer Science

Christian Albrechts University, Kiel

45

KRÖMKER Susanne

IWR – Interdisciplinary Center for

Scientific Computing, University

Heidelberg

35

41

KUCH Nora

Institut für Ur- und Frühgeschichte

und Vorderasiatische Archäologie

Ruprecht-Karls-Universität

Heidelberg

46

KULITZ Iman

IEMAR – Institute of Architectural

Science/Department Digital

Architecture and Planning

Vienna University of Technology

22

LANG-AUINGER Claudia

Institute for Studies of Ancient

Culture, Austrian Academy of

Sciences, Vienna

LÖCHT Joana van de

Institut für Ur- und Frühgeschichte

und Vorderasiatische Archäologie

Ruprecht-Karls-Universität

Heidelberg

46

MAIETTI Federica

DIAPReM – Department of

Architecture, University of Ferrara

10

MARA Hubert

IWR – Interdisciplinary Center for

Scientific Computing, University

Heidelberg

41

MEIER Thomas

Institut für Ur- und Frühgeschichte

und Vorderasiatische Archäologie

Ruprecht-Karls-Universität

Heidelberg

46

page

MOHAMED Fawzi

Institut für Chemie, Humboldt

Universität, Berlin

56

MONROY Antonio

IWR & Transcultural Studies,

University of Heidelberg

95

MÜLLER Hartmut

Institute for Spatial Information

and Surveying Technology, Uni-

versity of Applied Sciences Mainz

12

MÜLLER Noémi S.

Institute of Materials Science,

N.C.S.R. “Demokritos”, Aghia

Paraskevi

30

NEMES Paul

Faculty of Electronics, Telecom-

munications and Information

Technology, Technical University

of Cluj-Napoca

49

NGUONPHAN Pheakdey

IWR – Interdisciplinary Center for

Scientific Computing, University of

Heidelberg

45

88

NICCOLUCCI Franco

Science and Technology in

Archaeology Research Center, The

Cyprus Institute, Nicosia

53

NOBACK Andreas

Fachbereich Architektur

Technische Universität Darmstadt

28

NYS Karin

Mediterranean Archaeological

Research Institute, Vrije

Universiteit Brussel, Brussels

53

OMMER Bjoern

IWR & Transcultural Studies,

University of Heidelberg

95

Participants

111

page

PASKALEVA Galina

Institut für Architektur- und

Kunstgeschichte, Bauforschung

und Denkmalpflege

Vienna University of Technology

43

PECCHIOLI Laura

Institut für Geodäsie und

Geoinformationstechnik,

Technische Universität Berlin

56

PEREIRA John

Salzburg Research Forschungs-

gesellschaft m.b.H., Salzburg

59

PHAL Des

Information Technology Center,

Royal University of Phnom Penh

45

88

QUATEMBER Ursula

Österreichisches Archäologisches

Institut, Vienna

62

REINDEL Markus

German Archaeological Institute

Commission for Archaeology of

Non-European Cultures, Bonn

65

ROSSATO Luca

DIAPReM – Department of

Architecture, University of Ferrara

8

SANDAY John

Global Heritage Found, USA

SANTOPUOLI Nicola

Faculty of Architecture “Valle

Giulia”, La Sapienza, University of

Rome

10

SCHALLER Kurt

CHC – Research Group for Ar-

chaeometry and Cultural Heritage

Computing, University of Salzburg

24

85

SCHIEMANN Bernhard

Lehrstuhl Informatik 8 Künstliche

Intelligenz, Friedrich-Alexander-

Universität, Erlangen

32

page

SCHÖNLIEB Carola-Bibiane

DAMTP, Centre for Mathematical

Sciences, Cambridge

6

SCHULTES Kilian

Historisches Seminar, Zentrum für

Europäische Geschichts- und

Kulturwissenschaften, Ruprecht-

Karls-Universität Heidelberg

67

SDYKOV Murat

Public Foundation for Science and

Education «Akjol», Pedagogical

Institute, Saratov State University

Uralsk

68

SEDLAZECK Anne

Institute of Computer Science

Christian Albrechts University, Kiel

91

SHARON Gonen

The Archaeology Institute, The

Hebrew University, Jerusalem

26

SIART Christoph

Geographical Institute – Labora-

tory for Geomorphology and Geo-

ecology, University of Heidelberg

71

SINGATULIN Rustam

Public Foundation for Science and

Education «Akjol»

Pedagogical Institute of the

Saratov State University, Uralsk

68

74

SIOTTO Eliana

LIDA – Laboratory of Informatics

for Art Historical Research

University of Udine

77

SLIZEWSKI Astrid

Neanderthal-Museum, Mettmann

80

SMILANSKY Uzy

Department of Physics of Complex

Systems, The Weizmann Institute

Rehovot

26

Participants

112

page

SMIKT Oded

Department of Physics of Complex

Systems, The Weizmann Institute

Rehovot

26

SOTIROVA Kalina

Institute of Mathematics and

Informatics, Bulgarian Academy of

Sciences, Sofia

82

STRASSER Margareta

Center for Languages, University

of Salzburg

59

STRASSER Thomas

silbergrau Consulting Software

GmbH, Linz

59

THUSWALDNER Barbara

Österreichisches Archäologisches

Institut, Vienna

62

TRINKL Elisabeth

Institute for Studies of Ancient

Culture, Austrian Academy of

Sciences, Vienna

UHLIR Christian

CHC – Research Group for Ar-

chaeometry and Cultural Heritage

Computing, University of Salzburg

85

UNTERWURZACHER Michael

CHC – Research Group for Ar-

chaeometry and Cultural Heritage

Computing, University of Salzburg

85

VANUCCI Cristina

DIAPReM – Department of

Architecture, University of Ferrara

8

VAR Puthnith

Information Technology Center

Royal University of Phnom Penh

88

VIROLI Francesco

DIAPReM – Department of

Architecture, University of Ferrara

8

page

VISINTINI Domenico

Department of Georesources &

Territory, University of Udine

77

VLAICU Aurel

Faculty of Electronics, Telecom-

munications and Information

Technology, Technical University

of Cluj-Napoca

49

WAND Eku

Department of Communication

Design, Hochschule für Bildende

Künste Braunschweig

90

WINCKLER Michael J.

IWR – Interdisciplinary Center for

Scientific Computing, University of

Heidelberg

45

88

WITTUM Gabriel

Goethe-Center for Scientific

Computing, Goethe University,

Frankfurt am Main

41

WULFF Robert

Institute of Computer Science

Christian Albrechts University, Kiel

91

YAKOVENKO Olga

Public Foundation for Science and

Education «Akjol», Pedagogical

Institute, Saratov State University

Uralsk

74

YARLAGADDA Pradeep

IWR & Transcultural Studies

University of Heidelberg

95

YULE Paul

Institute of Prehistory and Near

Eastern Archaeology

Heidelberg University

99

ZAMBANINI Sebastian

Institute of Computer Aided

Automation, Vienna University of

Technology

101

113

ACKNOWLEDGMENTS

This publication was partially funded by the Heidelberg Graduate

School of Mathematical and Computational Methods for the Sciences

– DFG Graduate School 220.

Im Neuenheimer Feld 368

69120 Heidelberg, Germany

www.mathcomp.uni-heidelberg.de

The workshop was supported by Breuckmann GmbH.

Breuckmann GmbH

Torenstr. 14

88709 Meersburg

www.breuckmann.com

MONDAY, NOVEMBER 16TH

Welcome from the Organizers H. G. Bock11:15–12:30

Opening Remote Sensing and 3D-Modelling for the Documentation of Cultural

Heritage in Palpa, Peru

M. Reindel

Digital Geoarchaeology – an approach to reconstructing ancient

landscapes at the human-environmental interface

Ch. Siart

The Virtual and Physical Reconstruction of the Octagon and Hadrian’s

Temple in Ephesus

U. Quatember

Towards A Computer-based Understanding of Medieval Images P. Yarlagadda

Innovative mathematical techniques for image completion and

applications to art restoration

W. Baatz

14:00–15:40

Session 1

Scientifc

Computing I

Color Restoration in Cultural Heritage Images using Support Vector

Machines

P. Nemes

Heating efficiency of archaeological cooking pots: Computer models

and simulation of heat transfer

A. Hein

Automated GPU-based Surface Morphology Reconstruction of Volume

Data for Archeology

D. Jungblut

Boon and Bane of High Resolutions in 3D Cultural Heritage

Documentation

C. Hörr

16:10–17:30

Session 2

Scientifc

Computing II

An automatic method to determine the diameter of historical coins in

images

S. Zambanini

17:30–19:00

Welcome

Reception

Poster Session

TUESDAY, NOVEMBER 17TH

Data, Science, and the Media J. Broschart

Interactive Narratives for Exploring the Historical City of Salzburg J. Pereira

A multimedia museum application based upon a landscape embedded

digital 3D model of an ancient settlement

S. Boos

09:00–10:40

Session 3

Communicating

Sciences to the

Public Computing the “Holy Wisdom” O. Hauck

3D Reconstruction of Archaeological Trenches from Photographs R. Wulff

3D Reconstruction of Banteay Chhmar Temple for Google Earth P. Nguonphan

Practical experiences with a Low-cost Laser Scanner S. Kor

11:00–12:20

Session 4

3D-Reconstruction

and Modeling Computerized 3D modeling – A new approach to quantify post-

depositional damage in Paleolithic stone artifact

U. Smilansky

14:00–18:30 Guided Tour and Visit in Speyer

WEDNESDAY, NOVEMBER 18TH

Image-based techniques in Cultural Heritage modeling M. Sauerbier

GIS and Cultural Flows M. Arnold

A Collaborative GIS-based Cue Card System for the Humanities G. Christ

09:00-10:40

Session 5

GIS

Archaeological Information Systems I. Kulitz

“Archäologische Museen vernetzt“ – An Information System for the

Archaeological Museums in Bavaria

K. Schaller

Historic Quarries – the Database and Case Studies C. Uhlir

Das Projekt WissKI G. Hohmann

11:00-12:20

Session 6

Databases and

Information

Systems Artifact cataloging system as a rational translator from unstructured

textual input into multi-dimensional vector data

V. Klepo

12:20–12:40 Best Student Presentation Award

Closing Remarks

POSTER SESSION

The Angel’s cave - A database for the restoration and valorisation of

the San Michele Archangel site, Olevano sul Tusciano (Salerno, Italy)

M. Balzani, R. De Feo, C.

Vanucci, F. Viroli, L. Rossato

The 3D morphometric survey as efficient tool for documentation and

restoration in Pompeii: the research project of Via dell’Abbondanza

M. Balzani, F. Maietti, G. Galvani,

N. Santopuoli

MAG, an Italian XML application profile for submission and transfer of

metadata and digitised cultural content

P. Feliciati

Towards an Automated Texturing of Adaptive Meshes from 3D Laser

Scanning

Ch. Hoppe, S. Krömker

A new approach to the surveying of archaeological monuments in

academic teaching

Th. Meier, C. Casselmann, J. van

de Löcht, N. Kuch, H. Altenbach

The Hala Sultan Tekke site and its innovative documentation system F. Niccolucci, K. Nys, A. Felicetti,

S. Hermon

ISEE: retrieve information in Cultural Heritage navigating in 3D

environment

L. Pecchioli, M. Carrozzino, F.

Mohamed

IT of the PROJECT "VIRTUAL Ukek – Kyryk-Oba" M. Sdykov, R. Singatulin

IT in the reconstruction of ceramics R. Singatulin, O. Yakovenko

3D Texture Modeling of an important Cycle of Renaissance Frescoes in

Italy

E. Siotto, D. Visintini

NESPOS – Digitalizing Pleistocene People and Places E. Slizewski

Multimedia Galleries of Cultural Heritage – Piece of a Puzzle or the

Overall Picture?

K. Sotirova

One Land, One City, One Memory - Hidden in the Sand! – The

Famagusta Research Project

E. Wand

Laser scanning of a Mauryan Column Complex in Sisupalgarh, An

Early Historic Fortress in Coastal Orissa/India

P. Yule

DEMONSTRATION: Breuckmann GmbH, Meersburg