Discussion of session 9 readings: Ferdani 2019 & Vitale 2016

[gabrielbodard]

• Daniele Ferdani, Emanuel Demetrescu, Marco Cavalieri, Gloriana Pace, Sara Lenzi. 2019. “3D Modelling and Visualization in Field Archaeology. From Survey To Interpretation Of The Past Using Digital Technologies.” Groma 4 (2019). Available: http://dx.doi.org/10.12977/groma26
• Valeria Vitale. 2016. “Transparent, Multivocal, Cross-disciplinary: The Use of Linked Open Data and a Community-developed RDF Ontology to Document and Enrich 3D Visualisation for Cultural Heritage.” In: Bodard/Romanello, Digital Classics Outside the Echo-Chamber: Teaching, Knowledge Exchange & Public Engagement, Pp. 147–168. London: Ubiquity Press. Available: http://dx.doi.org/10.5334/bat.i

In addition to summarising the papers, please think about ways in which the content of both pieces affected your approach to the 3D modelling/visualisation exercise in this session.

[PAZHames]

Vitale - Transparent, Multivocal, Cross- disciplinary: The Use of Linked Open Data and a Community-developed RDF Ontology to Document and Enrich 3D Visualisation for Cultural Heritage

This article is about the problem of opacity in 3D technologies. 3D visualisations are largely presented to the public without documentation, meaning that the public cannot assess how accurate they are, as sources and references are entirely hidden. This is a particular problem in museums as it propagates the idea of the museum as an ‘ultimate authority’, suggesting that the visualisation presented is the only correct one. It also does not allow for alternative perspectives or conflicting evidence. Nor does it allow for the fact that a 3D visualisation is not a representation of the truth but that, like a photograph, it represents the subjective perspective of its author.

In 2006 the London Charter provided a set of guidelines for the use of 3D technologies, citing the importance of documentation but 10 years on the proportion of documented 3D visualisations is still very low.

Some attempts have been made:

• Work on the Villa of Livia at Prima Porta – but not clear how CD – ROM with the 3D visualisation actually linked to the archaeological research
• Digital Pompeii – documented, but the system could only deliver information when the source was part of that single digital repository
• The University of California Los Angeles in projects such as the Digital Roman Forum – but the provenance is often not declared and the methodology has not been followed systematically for all of the building components

All of these are heading in the right direction but none of them mention alternative hypotheses, lack of information or conflicting sources.

Proposal:

Developing a new documentation standard to make the process time and cost effective. Vitale suggests creating a dedicated resource description framework ontology using Linked Open Data. This would make the documentation comparable and not idiosyncratic. It can be read by both humans and machines so connections can be automatically identified and presented to users. It was suggested that the process start by looking at museum ontologies, but none of these were found to be specific enough. It is important that the ontology be written by a group, as creating an ontology is a knowledge representation process. If it was based on only one researcher’s view of the world, it would be of limited use to others. The suggested ontology (offered here as proof of concept) is called SCOTCH (Semantic Collaborative Ontology for Three-dimensional visualisation of Cultural Heritage).

URIs:

To express something as linked open data, it must be divided down into smaller units and given a Unique Resource Identifier (URI). These URIs should be as neutral as possible and without any attributed taxonomies, such as ‘temple’ This breaking down aids debate over specific elements of a 3D visualisation. Each part should have the name/function, author and hopefully bibliography. It can have multiple annotations to show varying interpretations.

Sources:

The danger of expressing connections between elements through ‘certainty’ is also discussed, as the most certain is not necessarily the most accurate, in archaeology work is often ongoing, and no work is every truly ‘finished’. Instead of using levels of ‘certainty’, it is proposed that the ontology refer to the type of source, such as expert knowledge or direct observation. It is stressed that all types of source are acceptable, as 3D visualisation does not represent the actual artefact, but the knowledge of the author about the artefact.

Benefits of this ontology:

• Transparency
• Repeatability
• Peer review
• Citation
• Can make 3D research multidisciplinary, as the documentation will be comparable across all industries.

Outside academia it could be used for:

• Museums
  • To show behind the scenes and conflicting theories
  • The public could add their own annotations, which has real informative value, especially as some might be last witnesses to otherwise lost cultural heritage
  • 3D printing – enhanced understanding, helpful for the visually impaired and of use to smaller museums
  • Also useful to link data between repositories in different ways, meaning the possibility of more discoveries and isolating gaps in the museum’s own information
• Education – to emphasise the fact that all we know about the past is hypothetical
• Artefacts and building restoration – when measurements are taken over time or with different technology it can show changes over time and the possible biases of different technologies
• Archaeological sites
• Geological or hydrological surveys – emphasising the interdisciplinary potential
• To show how human activities have changed a landscape
• To map underground areas of cities and help with transportation links

The ontology and its potential development do however run into some issues:

• Still experimental
• Complicates issue of intellectual ownership
• Copyright issues - whether people are willing to share their work
• Would mean 3D tech researchers can be assessed - do they want that?
• Whether the community is actually interested in expanding the ontology - would require time and resources (even if Figuring out what the most appropriate technology to use is - author thinks OpenGL

I have used a lot of bullet points as the article contains a lot of lists and I thought this might be the clearest way to present the information – I hope this works for you guys!

I thought this was a really excellent and very ambitious article. One of the smaller points which I really like was the potential to use 3D visualisation and this ontology as a way of mapping the effects which humans have on an area over time. Although the project is still in its nascent stages, I was really interested in its potential for presenting the effects of climate change on our world in an engaging and impactful way. I also liked the idea of potentially allowing the public to add annotations and the way it tied into Majd Al-Shihabi’s presentation from a few weeks ago (that one really stuck with me in case you couldn’t tell!) re last witnesses to otherwise lost cultural heritage. Please do let me know if any of this is not clear enough and looking forward to hearing what you think!

[nicolealexandra33]

With 3D visualizations, I wonder about the increased complexities surrounding 3D facial reconstructions-- in this situation, I guess there is a bit less controversy surrounding who is the person reconstructing, as creating facial reconstructions requires someone who is trained in forensics. However, there have still been quite a few issues surrounding the 'whitewashing' of historical figures' 3D models, particularly when it is hard to determine their hair/skin/eye colour, as the forensic reconstruction can provide you with a facial structure and features, but not of the above features' colours. With greater advances in DNA analysis, it may be possible going forward to determine an individual's hair/skin/eye colour, but until then, I can see that there will still be issues in 'finishing' these 3D facial reconstructions, as DNA or other artwork depicting the individual is not available.

[K-Doering]

Daniele Ferdani, Emanuel Demetrescu, Marco Cavalieri, Gloriana Pace, Sara Lenzi. 2019. “3D Modelling and Visualization in Field Archaeology. From Survey To Interpretation Of The Past Using Digital Technologies.”

This article is about the merging of hard sciences with the humanities through new computer-based research technologies, which interpret and visualize archeological data. The case study uses the roman villa of Aiano.

INTRODUCTION
The application of 3D modeling tools is typically considered communicative and interpretative rather than a part of a systematic scientific process. Archeological data collected through excavation is used in combination with reconstructive archeological interpretation.

DEALING WITH THE RECONSTRUCTION OF THE PAST
Computer-based visualization to show theoretical reconstruction has been a part of archeology since the late 1980s. With technological progress the capabilities of computer-based 'virtual archeology' and reconstruction have increased dramatically. An issue with digital reconstruction is that the purpose, as was stated in the Youtube seminar, is to lead to a better understanding of the past and not to be taken as 'truth' or 'statements of reality.' The authors note that all virtual reconstructions should declare the level of authenticity to ensure that the user can discern what is real from what is interpreted. Most reconstructions do not declare such levels of authenticity, usually due to the complexity of sources and contexts.

The first applications of virtual reality for 3D visualization of archeological data had several weaknesses:

• Lack of consistent projects aimed at finding precise answers in the archeological data, lack of focus on scientific research
• Reconstructions were presented as reality without taking into account alternative hypotheses
• Lack of interdisciplinary professionals

In addition, the first applications had the following strengths:

• Took advantage of new medium to gather new insights
• Good way to test complex hypotheses and visually absorb complex data
• Facilitates understanding information

THE INTELLECTUAL TRANSPARENCY
Due to the debates on the uses of virtual reconstruction, regulative practices and guidelines have been implemented to standardize and regulate the use of such technologies (e.g. London Charter and the Seville Principles). The London Charter 'defines a set of principle of computer-based visualization to ensure the intellectual and technical integrity, reliability, documentation, sustainability and accessibility.' The particular importance of the Charter is that it highlights not only the documentation of sources for the virtual reconstruction but the interpretation made.

THE CASE STUDY
A villa was discovered in Aiano (near San Gimignano) and the oldest portions dated from the late 3rd/early 4th century AD. The villa's construction and decorative details were altered/recycled in the late 5th/early 6th century AD. Only the base of the walls are preserved along with the decorative elements (found outside of their original context due to rearranging and recycling of materials).

METHODS INTO PRACTICE
Creating hypothetical models is a challenging task and involves a philological approach based on derived data and the expertise of several professionals (e.g. archeologists, historians, architects, etc.). The quality of the model is dependent on the quality of the data used and the analysis of said data. It is often necessary to be dependent on additional materials such as historical documents or comparative studies.

RESEARCH PROBLEMS AND EXPECTATIONS

• Stylistic and historical comparisons are difficult to procure due to the decorative elements being found out of their original context
• Could have better organized the structure of archeological data to improve analysis
• Met issues in integration of gaps, degree of reliability of hypothetical reconstruction, and with 'transparency' in the representation of the data

DIGITIZATION OF THE ARCHEOLOGICAL SITE

1. Geometric acquisition of the archeological sites in current state of preservation using Dense Image Matching (DIM) techniques -- particular focus on the trefoil hall
2. Topographic and photogrammetry survey -- images don't provide metric data so further topographic information needed to be gathered using Ground Control Points (GCP) and an Unmanned Aerial Vehicle (UAV)
3. Images were processed with Agisoft Photoscan -- computed by algorithm of Structure From Motion (SFM). The features extracted were associated to 3D points and constructed into a sparse point cloud and computed by algorithms of Dense Stereo Matching (DSM). The points were then triangulated to obtain mesh models with textures. Topology of the mesh was optimized using CloudCompare.
4. Topographic network of points was appended to the model for scale and geo-referencing

SOURCE COLLECTION AND MANAGEMENT
As stated previously, the completeness and the reliability of the reconstruction is based on management and completeness of the data being used. The team of researchers collected metric survey data, photos and sheets of collapsed fragments and structures, photos of archeological finds, archeological reports and sketches, existing structures and Roman modules, and comparisons to still preserved buildings.

There is always a risk of creating models that could have never existed -- structural and functional comparisons were used in the case study techniques.

INTERPRETATION AND 3D MODELING
All of the data used in the case study was cross-referenced and then simulated using 3D modeling techniques. The reconstruction gave the 'archeological remains architectural completeness.' Sources were inefficient to reconstruct the entire building so simulations were crucial. All elements were simulated to give as complete of a rendering as possible. These renderings made further hypotheses possible (e.g. suppose the type of wood used in structure/heights of the walls) and less likely (e.g. presence of vaulted ceilings).

The reconstruction procedures were registered using the Extended Matrix approach. This method records not only archeological stratifications but also hypothetical reconstructions. Sources were also connected to paradata in Blender and a proxy (a prototype which allows the analysis of matrices and excavation information in the reconstructive process). Polygonal modeling was used to complete the realistic visualization of the model (e.g. details and materials used to create the structure(s)).

CONCLUSION
The use of the 3D modeling in the archeological context can improve interpretation through simulation and analysis. Modeling enables researchers to test their interpretative hypotheses and 'transparency' in the presentation of the data is crucial.

This article was certainly an interesting read. I am particularly drawn to the idea of standardizing the ways in which virtual reconstruction can and should be presented. It was not even a question that crossed my mind when seeing a 3D reconstruction of a site, like in a documentary. I was of course aware that it is a reconstruction but I think that making the line evident of where archeological data ends and interpretation begins is absolutely crucial. If we have strict guidelines for the preservation of standing historic sites, we should also have such standards for modeling, especially if the model is to be shared with the public.

[ChantalvanEgdom]

Great summaries both! I think it's great that there is more thought about decisions and even skills of 3D modellers. When doing the exercise, I found myself sometimes setteling for the next best maesurement as I couldn't quite seem to pinpoint the exact spot, and when this happens in models that will actually be used, beit in a museum setting, academia or anything else published, I'd like to be aware of those choices. However I do see the strain it may put on people wanting to publish something but being held back by the higher standards, so I guess the extent of using such a standard would depend on the aim of the model? Also, great point by @nicolealexandra33 ! Choices made by an individual working on a 3D model, could have huge consequences on people's perception of for example skin colour and result in white-washing, as you pointed out. I guess having certain standards, or at least providing a report of the choices that were made by the modeller would be an improvement? What do you guys think?

[K-Doering]

...However I do see the strain it may put on people wanting to publish something but being held back by the higher standards, so I guess the extent of using such a standard would depend on the aim of the model? Also, great point by @nicolealexandra33 ! Choices made by an individual working on a 3D model, could have huge consequences on people's perception of for example skin colour and result in white-washing, as you pointed out. I guess having certain standards, or at least providing a report of the choices that were made by the modeller would be an improvement? What do you guys think?

Great points and questions, both! I think that providing a report of the choices made (upon request even) would be extremely beneficial, particularly for any models that would be used in somewhere like a museum or in academia. As mentioned in the Vitale article, I definitely could see the issues with researchers work being thrown out because of the choices made without references -- however, I do believe that the point of these types of models are not to be exact replicas of ancient sources but aids in expressing hypotheses. I do think if this is evident (something as simple as a disclaimer) then I think a model can be put forth to the public without a formal report but reporting should be a part of the process on the researcher/developer end if they want any sort of credibility for their work and choices.

I did appreciate that the article expressed that the way in which models are created now are not using the scientific method and an important aspect of experimentation is replication but I do think that the models are more of an art than a science, and that's okay.

[FabioDFernandes]

Thank you both for your excellent and super clear summaries! I agree @K-Doering that we often just take (or I have done) 3D reconstructive models of ancient sites at face value without ever considering what exactly is tangible and verified, and what is still hypothesis. It might give the very outward impression to many that we have a very complete image of historic sites that today are no less than the typical leftover wall bases, but in reality we are far from this, and these are reconstructions based on our still fragmentary understanding. I agree that a clear line should be drawn, even in non-academic documentaries/publications, etc. I suppose it really is a matter of 'transparency' and intellectual honesty. But it is nonetheless a very useful technology in this context, to be able to virtually 'reconstruct' and visualise, rather than trying to picture with our own minds the data we have or theorise on.

@PAZHames I also agree regarding the potential to use 3D visualisation to map human effects on an area over time. This is such a recurring theme in this course, but yet another occasion where public annotations are welcome, especially when they might be useful in demonstrating such effects on cultural heritage (like Al-Shihabi's really touching project on the altering and vanishing Palestinian presence in certain areas), but with global warming too and more local effects, like you said - what two great examples. I've not exactly added anything new on my part, but I would love to learn more about it. It's a very interesting paper overall, and the point that museums present 3D visualisations to the public without documentation and thus propagate themselves as an 'ultimate authority' whilst not allowing for alternative perspectives or conflicting evidence, ties quite nicely into the other article, which discusses how reconstructive models do not highlight their interpretational parts and nature. Wider transparency to the public and other specialists would serve to make information more accessible, but also be potentially very useful in several contexts. Of course, a random person might not be able to contribute effectively to a Pompeii-related 3D visualisation, but they could offer perspective on objects linked to their own heritage or influenced by their own potential disciplines, and other specialists who are not part of such studies could also fill in gaps (falls into the peer review benefit).

[ChantalvanEgdom]

@K-Doering , I think you're right, especially "that the models are more of an art than a science, and that's okay." Standards and reports and their amount of detail should depend on the usage of the model.