Dear Mar Shiprim readers, we are very happy to introduce to you a project which, we believe, will be of great interest to many of us! The “Virtual Cuneiform Tablet Reconstruction (VCTR) Project” is an international collaboration inspired by the ambition to support virtual access to cuneiform artefacts and to reconstruct cuneiform tablets by joining virtual fragments together. Enjoy the reading!

 

The “Virtual Cuneiform Tablet Reconstruction Project” (VCTR)

Sandra Woolley (Keele University, Staffordshire, UK), Tim Collins (Manchester Metropolitan University, Manchester, UK) and Erlend Gehlken (Goethe-Universität, Frankfurt am Main, Germany)

We met at the University of Birmingham (UK), inspired by the idea of facing new technological challenges. We wanted to develop an algorithm with the help of which hopelessly destroyed artefacts could be virtually reassembled with the help of 3D scans without further “human intervention”. As a starting point for our investigations we took the 5400 fragments of the so-called Eanna archive from the 12th and 13th excavation campaigns in Uruk from the time of Nebuchadnezzar II. Half of the fragments are in the National Museum in Baghdad, the other half are on permanent loan from the German Archaeological Institute (DAI) to Heidelberg University. Our approach is based on looking for fragments that “physically” fit together (form a “join”). A (more recent) approach used by colleagues proceeds from “joining” texts with each other via transcriptions (i.e. quasi “philologically”). This approach is vividly described by Enrique Jiménez in Mar Shiprim, but it is (essentially) only usable with literary texts; business texts fall outside the framework because they are (generally) unique, although in many cases they share a common “basic form”. Combining the two approaches would be an ideal way to join texts. A photograph may illustrate this:

Figure 1: K 2271 + ? + ? (+) K 9515 + K 12355 (+) K 18437 (kind permission of the Trustees of the British Museum)

In the join shown (photograph published with kind permission of the British Museum, London), a piece of a rikis gerri tablet (a “guideline tablet” to the Babylonian omen series Enūma Anu Enlil), the direct joins K 9515 + K 12355 (upper joined piece in the photograph) and K 2271 + ? + ? (the fragments were already joined when they were handed over to us) could be found easily with the methods of the VCTR project, the two indirect joins (connections of the upper and the lower piece with the middle fragment (K 2271+) required the approach described by E. Jiménez or – as in this case – the expertise of the person who took the photograph after arranging the fragments accordingly (the joins were made manually a longer time ago).

The basics of the VCTR approach have already been published in several publications (for example, in 2014 in a NABU article entitled ‘Automated Joining of Cuneiform Tablet Fragments ’ and in an IEEE conference paper in ‘Computer-Assisted Reconstruction of Virtual Fragmented Cuneiform Tablets’. There will only be a short summary to follow. In the meantime, the project has grown far beyond its original narrowly defined boundaries. These new developments, made mostly at Keele and Manchester Metropolitan Universities, will form the main part of this paper.

1. The VCTR Project – Aims and Technologies

Figure 2: The Virtual Cuneiform Tablet Reconstruction Project Website: https://virtualcuneiform.org/

The VCTR project website, virtualcuneiform.org, provides information about the project, underpinning research, and achievements in the virtual reconstruction of fragmented cuneiform tablets. The site also showcases a range of 3D interactions and provides links to publications.

A significant aim of the VCTR project was the automated virtual reconstruction of fragmented tablets. Automated reconstruction methods had been demonstrated for some reconstruction applications, e.g., pot sherds, skull fragments and figurines, but they tended to assume a near-perfect fit between the fragments. To account for the erosion usually evident on cuneiform tablet fragments, a novel automated reconstruction algorithm was devised. The reconstruction algorithm was designed to automatically establish whether a pair of 3D-scanned virtual tablet fragments join and to determine the required orientations of the fragments to achieve the best fit. Figure 3 shows the first successful demonstration of the algorithm using a pair of tablet fragments originating from Uruk.

Figure 3: The First Automated Join Achieved for Late-Babylonian Uruk Fragments

The initial tests of the automated reconstruction algorithm used 3D scans of tablet fragments that had already been joined but not yet glued together. To demonstrate the further potential of the technique, the next test was a long-distance virtual join. We selected the third tablet of Ipiq-Aya’s Old Babylonian inscription of the Atrahasis epic, most parts of which are held in the British Museum whilst another significant fragment – believed to be a likely join – is held in the Musée d’Art et d’Histoire, Geneva.

Figure 4: The First Long-distance Automated Join for Virtual Fragments of the Atrahasis ‘C’ Tablet

Using our custom 3D acquisition system, separate virtual models were generated in London and Geneva and the join proven using the automated joining algorithm, despite the actual fragments remaining separated by approximately 1000 km (see figure 4). An animated visualisation of the virtual Reconstruction of the Atrahasis ‘C’ Tablet can be viewed at https://www.youtube.com/watch?v=lSzOY36Z-FA

2. Virtual and Augmented Reality Interfaces

One of the ambitions of the VCTR project has been the development of tools and interactive virtual environments to support 3D viewing and interaction with virtual cuneiform tablets. Our first iterations were computer-based and designed for virtual collaborative tablet reconstruction. Our first online viewer, in 2017, brought this facility to standard web browsers allowing anyone with internet access to view and manipulate virtual cuneiform tablets. Improvements followed in the functionality and in the rendering options; the evolution of the interactive viewer can be seen here.

Figure 5: The Prototype Android Augmented Reality (AR) Museum Smartphone App

Augmented Reality (AR) apps for both Android and iOS smartphones have also been prototyped as interactive interfaces to cuneiform tablets and other artefacts. Our aim was to demonstrate the potential for an Augmented Reality Museum, enabling users to create and curate their own virtual interactive artefact exhibitions at home. The interface for the Augmented Reality app was designed for educational use, with pop-up information about the artefacts available, as well as alternative ‘wireframe’ and ‘point-cloud’ viewing options to illustrate how the virtual 3D models were created (see figure 5).

3. Google Summer of Code

3D artefact viewer interfaces should support the interactions needed by users. For example, viewer interfaces that fix artefacts upright may work well for figurines or other standing artefacts that users would not need to view upside down, but such viewers are not suited for cuneiform tablets whose script wraps vertically from the obverse to the reverse side.

The VCTR 3D viewer interfaces were designed to support rotation, joining, lighting and other interactions necessary for viewing and interacting with cuneiform tablets. One of the evolved versions of the web-based VCTR 3D cuneiform tablet viewer was integrated into the new Cuneiform Digital Library Initiative (CDLI) platform via the CDLI Google Summer of Code projects led by Émilie Pagé-Perron. The 2021 student intern, Mustafa Dhar, successfully integrated the viewer with VCTR mentorship.

4. Public Engagement

One aspect of virtualising 3D artefacts and making them available via on-line interactive interfaces, is that the artefacts are now made available to wider audiences across the world. To encourage the educational potential of our interactive viewers, the VCTR provides information for schools including an overview of cuneiform writing, the technology behind virtual 3D models and a 3D interaction join challenge that challenges users to join a pair of virtual fragments themselves using the online 3D reconstruction interface. Website usage data shows us that this resource has proven useful to educators internationally.

5. Résumé and Plans for the Future

The VCTR project has made advances in the use of 3D scanned virtual cuneiform tablets for automated reconstruction and for widening access via Augmented Reality apps and online 3D viewers including an implementation now integrated into the CDLI. Future plans include further enhancements to the 3D viewers including annotations and measurement tools.

Categories: Mar Shiprim