International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
will remain a major research area, it is our opinion that
documentation has been neglected. Most scholars assume that
virtualised models are conformed to the original ones within the
precision of the apparatus utilised to virtualise them. It has
been shown that although a model might be precise, it is not
necessarily accurate [3]. Precision is related to the
reproducibility of the measured data while accuracy is related to
the conformity of the measured data with the original model. It
is difficult in an extra-laboratory situation to determine if the
measured data are accurate or not. Unfortunately, this is the
most common situation. For that reason, in addition to the
virtualised model, additional information should be provided.
The provided information should at least include, the type of
apparatus utilised to virtualise the artefact, resolution, precision,
time-stamped calibration, time-stamped raw data, time-stamped
temperature and humidity, etc. The availability of such data
would assist the scholar to assert the validity of the measured
data.
Indexation turns out to be important when the number of
virtualised artefact becomes significant. Our aptitude to retrieve
pertinent information is primarily determined by the quality of
our indexes. An index can be a textual description of the
artefact, a set of measurements, a set of keywords and metadata
or, an innovative technique like composition and shape
description: the so-called content-based approach. While some
fields like archaeology have developed a systematic procedure
for indexation, one has to recognise that it is not the case for
most fields. Even worst, in most cases, there is no information
at all associated with the artefacts. A content-based approach
can provide a partial solution to this problem since it
automatically indexes a collection. The content-based indexes
provide a description of the artefacts in terms of composition
and shape. A search engine then provides a fast and efficient
access to the collection. The content-based approach does not
replace classical indexation but it complements it and provides
an alternative 1f the later is absent.
As we have seen earlier, characterisation is intimately related to
retrieval. The combination of classical documentation with
content-based indexes is of particular interest. Classical
information can act as a filter on the collection, which then can
be searched with a content-based approach. Some results in
relation with this approach are reported in [12]. Once the
artefacts of interests have been retrieved, that must be
visualised. We have presented a cost-effective stereo visual
system for the visualisation. of complex three-dimensional
artefacts and scenes. One of the advantages of this system is
that it can be rapidly deployed on any working site and may
therefore be utilised as a terrain and operational device.
5. CONCLUSIONS
The creation of virtual collections of cultural heritage sites
provides both experts and novices with unrestricted access to
these places of interest. The stereo environment, as described
here, provides a cost-effective system to visualize, search and
further explore such virtualised collections. The system is
mobile, scalable and is able to visualise large amounts of
complex three-dimensional data, as illustrated by the use thereof
to visualise a number of cultural heritage sites in Italy.
Experimental results show that use of content-based retrieval,
together with traditional information retrieval techniques, offer
the user the ability to accurately characterise virtual collections.
Here, the “query by example” or “query by prototype” paradigm
allows for the clustering of similar pictures and three-
dimensional artefacts. Through the use of this technology,
novel insights into previously restricted historical and
archaeological sites can be obtained through comparative
studies, virtual model documentation and visualisation.
6. REFERENCES
1. Kuman, S. et al, 2003. Digital Preservation of Ancient
Cuneiform Tablets Using 3D Scanning. In: Fourth
International Conference on 3-D Digital Imaging and
Modeling, Ottawa, Canada, pp. 326-333.
2. El-Hakim, S. F., Beraldin, J.-A. et al., 2003. Effective 3D
Modeling of Heritage Sites. In: Fourth International
Conference on 3-D Digital Imaging and Modeling, Ottawa,
Canada, pp. 302-309.
3 Guidi, G. et al, 2003. Accuracy Verification and
Enhancement in 3D Modeling: Application to Donatello’s
Maddalena. In: Fourth Internaional Conference on 3-D Digital
Imaging and Modeling, Ottawa, Canada, pp. 334-341.
4. Beraldin, J.-A., Valzano, V. et oL, 2003. Virtualizing a
Byzantine crypt: challenge and ;impzct. In: Videometrics VII,
Santa Clara, USA, pp.137-147.
5. Gaiani, M., 2000. Strategic di rappresentazione digitale:
modelli per la storia e la conservazione dei beni architettonici e
ambientali. Centro di Ricerche Informatiche per i Beni
Culturali, 10, pp. 47-69.
6. Hakim, S. L., Gonzo, L. et al., 2003. Visualization of Highly
Textured Surfaces. In: 4th International Symposium of Virtual
Reality, Archaeology and Intelligent Cultural | Heritage,
Brighton, United Kingdom, pp. 1-9.
7. Tsekeridou, S. and Pitas, I, 1999. Audio-Visual Content
Analysis for Content-Based Video Indexing. In: [EEE
Multimedia Systems, Florence, Italy, Vol. 1, pp. 667-672.
8. Paquet, E., “Content-based Indexing and Retrieval of 3D and
Image Databases", Ottawa, Canada. http://w ww.cleopatra.nrc.ca
(accessed 7 April 2004)
9. Paquet, E. and Rioux, M., 1999. Nefertiti: a query by content
system for three-dimensional model and image databases
management. /mage and Vision Computing, 17, pp. 157-166.
10. Han, J. and Kamber M., 2001. Data Mining: Concepts and
Techniques. Morgan Kaufman Publishers, Boston.
11. Paquet, E. and Rioux, M., 2003. Anthropometric Visual
Data Mining: a Content-based Approach. In: International
Ergonomics Association XVth Triennial Congress, Seoul,
Korea, pp. 78-91.
KE
ABS
In tl
ca 2
belo
the '
this
enti
The
Secc
in tl
the
acro
Stra
Nor
mee
Bud
sedi
260
car
Bud
Kak
200
In 2
the
Ban
[htt
Con:
acti
imn
frag
mur
