An Ontology for 3D Visualization in Cultural Heritage
Valeria Vitale (King's College London)
Digital Classicist London & Institute of Classical Studies seminar 2013
Friday June 14th at 16:30, in Room G37, Senate House, Malet Street, London WC1E 7HU
To date, 3D computer graphics and modelling techniques have been used in the study of the ancient world mainly as a means to display traditional research. The value of these digital techniques has been often assessed merely on the degree of graphic aesthetic quality.
The pursuit of “photorealism” has proven ineffective in engaging the audience but also scientifically misleading, as it suggests that is possible to reproduce an artefact or scene “exactly as it was” in the past.
Behind every scholarly 3D visualisation is a thorough study of excavation records, iconographic documentation, ancient literary sources, artistic canons and precedents. However, this valuable research (that may lead to new discoveries in the field) is not always detectable in the final visual outcome.
The London Charter for the Computer-based Visualisation of Cultural Heritage made a huge step forward in the regulation of scholarly 3D visualisation—prescribing that researchers’ choices and motivation must all be documented. No 3D model could be considered a scholarly resource if its research method was not “transparent”.
The London Charter presents methodological guidelines for recording this data, but does not go as far as to offer a formal framework in which to place this information; each modeller is left to simply follow their own style. Moreover, the clients who commissioned the 3D model (such as museums or other cultural institutions) are frequently more interested in the final product than in the rationale which is often completely overlooked and not circulated (or, in the worst case, dropped from the budget line altogether).
Since there are programming languages that enable 3D environments to successfully interact with html, I propose that it would be useful to create one or more ontologies to standardise the verbal component of the documentation, embedding it in the 3D model itself.
More specifically, I would suggest
1. A descriptive ontology
In a scenario where software and format may change quickly, a descriptive ontology could be helpful to rebuild data and metadata if the visual component was not readable anymore, enhancing accessibility, sustainability and longevity of the information. Its main aims would be:
* Describe the digital object
In the first place, it would be possible to assign a specific URI to each element of a 3D digital visualisation (@prefix “obj”: <http://hypothetical3donthology.kcl.ac.uk/objects/>).
Then, each digital object could be associated to its creator(s); the software used to generate the file; the different formats in which it is available.
A pre-existing ontology or formalised thesaurus of ancient art (such as CIDOC CRM, something based on the Getty Thesaurus of Art and Architecture, or the forthcoming UK Archaeological Data Service ontology) could be used, and adapted, to define the objects.
For example obj:001 rdf:type art:shaft.
Where @prefix “art”: <http://hypothetical3donthology.kcl.ac.uk/ArtVocabs/>
* Describe object’s relationships with other 3D digital objects
Through a dedicated namespace (@prefix “tdvo”: <http://hypothetical3donthology.kcl.ac.uk/threedvisontology/>) it could also be possible to state and describe properties, values and relationships of the 3D digital objects such as the relationship between a 3D digital object and the file it belongs to (obj:001 tdvo:isPartof obj:3Dfile.max) or the relationships between different objects within the same file (obj:001 rdf:type art:shaft. obj:010 rdf:type art:column. obj:001 tdvo:isPartof obj:010).
* Describe 3D digital object’s relationships with its physical referent
Through Pleiades, the 3D digital object could be linked to the place where the visualised building (or other referent) is located. For example, for a file “3Dfile.max” visualising the Odeon in Aphrodisias we could have: obj:3Dfile.max tdvo:hasLocation pleiades:638753.
In some cases, it would be possible to use Pleiades to link the digital visualisation even more specifically to the exact position of the building.
For the same example: obj:3Dfile.max tdvo:hasLocation pleiades:638753/odeon. Different 3D visualisations could be connected to the physical building and be available alongside the photographic documentation linked to Pleiades via Flickr. Using Pleiades vocabulary:
obj:3Dfile.max gawd:depicts pleiades:638753 or, (potentially) more precisely, obj:3Dfile.max gawd:depicts pleiades:638753/odeon
A methodological ontology
to define and represent:
* the level of speculation involved in the creation of each element.
For example obj:001 tdvo:hasCertainty tdvo:certainty6
where the level of certainty from 6 (maximum) to 0 (minimum) would be defined as follows:
tdvo:c6 rdfs:label “Certainty 6”; rdfs:comment “the ancient element is still in situ, and its dimensions and position can be measured”.
tdvo:c5 rdfs:label “Certainty 5”; rdfs:comment “the ancient element is not in situ but it has been visually documented in the past and the documentation is still available”.
tdvo:c4 rdfs:label “Certainty 4”; rdfs:comment “the ancient element is not in situ but it can be geometrically derived from the surviving elements”.
tdvo:c3 rdfs:label “Certainty 3”; rdfs:comment “the ancient element is not in situ but it can be visualised according to well accepted standards and precedents”.
tdvo:c2 rdfs:label “Certainty 2”; rdfs:comment “the element is not in situ but it can be visualised according to the modeller’s experience, knowledge, intuition”.
tdvo:c1 rdfs:label “Certainty 1”; rdfs:comment “the element is not in situ and it has been added for communicative purposes”.
tdvo:c0 rdfs:label “Certainty 0”; rdfs:comment “the element has not been created for scholarly purpose and does not aim to historical accuracy. However, some characteristics of an original referent can still be recognised”.
* the relationship between the 3D digital visualisation, its sources and referents.
- tdvo:isBasedOn rdfs:label “is based on”; rdfs:comment “the shape, dimensions or decoration of the element is based on visual or written information contained in a relevant document describing established practices, standards and rules”.
- The object of the predicate could be traditional bibliographical references and/or the digital URI of the source and/or the URL of a digital edition of the source such as the ones available on open digital archives (obj:001 val:isBasedOn dbpedia:De_architectura).
- tdvo:hasEvidenceIn rdfs:label “has evidence in”; rdfs:comment “the 3D element can be compared with specific verbal or visual evidence such as video/photographic documentation or official excavation records”.
- The object of this predicate would be archive numbers or bibliographical references identifying physical documents or artefacts, and/or URIs of digital reproductions of them, available on digital databases such as Arachne or CLAROS. For example, if obj 002 was an element of the 3D visualisation of the Basilica in Pompeii:
obj:002 tdvo:hasEvidenceIn <http://arachne.uni-koeln.de/item/marbilder/2015507>
- tdvo: isMentionedIn rdfs:label “is mentioned in”; rdfs:comment “the visualised building (or part of it) is mentioned in a ancient (or modern) text”.
- tdvo: isDescribedIn rdfs:label “is described in”; rdfs:comment “the visualised building (or part of it) is described in a ancient (or modern) text”.
- The latter predicates could link to bibliographical references and/or to the digital version of ancient texts such as the ones available through Perseus Project.
The two ontologies would be useful to:
- constrain the documentation, creating a standard,
- speed up the recording process thus reducing time/cost and making the documentation more likely to be retained in a project,
- allow and encourage comparison of different visualisations and interpretations of ancient heritage,
- contribute to transform 3D visualisation from univocal display of traditional research to a collaborative virtual environment that can be shared and implemented by different scholars,
- allow citations and re-use of the visualisation of entire buildings or single elements and peer-review,
- making 3D visualisations (human- and machine-) searchable, connecting them with the literary and historical sources that mention the visualised artefact or building.
In this way, a user interested in the visualisations could find and read the sources that are behind the modeller’s choices and judge if they are grounded or not. At the same time, a user interested in the sources could find (different) digital visualisations of how the artefact or building described might have looked like. Verbal and visual (two- and three-dimensional) resources could be accessed together and studied with an integrated approach, enhancing the comprehension of the past.
The seminar will be followed by wine and refreshments.