Report | 18. December 2024

»Give Your Data More Meaning in All Dimensions« Final report

By Dr. Lozana Rossenova

On 27th November, 2024, the final workshop and networking event of the “Give your data more meaning in all dimensions” programme took place in Berlin, at the Staatsbibliothek offices. Project representatives from five of the participating projects joined NFDI4Culture TA1 team members, as well as collaborators from the Kompakkt Consortium from the University of Cologne, alongside colleagues with relevant expertise from the SBB and the TIB, in person and online.  The projects participating in the programme presented their cases for using 3D models in their collections presentation and research, and for connecting the models with enriched metadata and annotations. They shared experiences from their ongoing research within their own institutions and from working with the Semantic Kompakkt platform. The purpose of the event was to gather feedback, use-cases and ideas from all participants regarding the future roadmap of the Semantic Kompakkt project, as well as to exchange experience and best practices in developing, processing and doing research with 3D data. Links to the individual presentations can be found below.

Learning from shared experiences across cultural domains

In the morning, we heard from Dr Sabine de Günther on her research into “Contextualized garments in the digital realm” from the projects Restaging Fashion at FH Potsdam and Virtual Couture at Kunstgewerbemuseum Berlin. We learned much about the unique challenges posed by capturing different varieties of textures with 3D scanning and modeling techniques, and the possibilities to create engaging web-based stories around the models with animation and movement. Learning points for the Semantic Kompakkt team included the need to consider animation within annotations, and the need for post-processing within the Semantic Kompakkt client to enable upload of larger and more detailed models, as well. 

The team from the ETH Library, represented by Roberta Spano, presented their collection of 3D models of scientific instruments and teaching aids. They discussed their experiences comparing 3 different viewer environments for 3D models on the web: Sketchfab, Morphosource, and Semantic Kompakkt. Thanks to their experimentation and possibility to clearly compare the performance of the same models in each environment, the Semantic Kompakkt team received valuable feedback throughout the course of the programme and were able to improve various features in the software. A key issue that was raised with this project was the need for an easy pipeline that allows transferring objects from environments like Sketchfab directly to Semantic Kompakkt. 

We also heard from Luca Junge from the German Maritime Museum, who discussed the preparation and upload of 3D scans of boat models from the museum collection and a collection of smaller decorative objects from the royal yacht Hohenzollern, also housed at the museum. The museum had developed an effective post-processing pipeline involving multiple tools to clean up the mesh captures and make them as small as possible for web applications. The typical order of their process was as follows:

1. Using MeshLab (for closing holes, reducing the number of vertices);

2. Using Blender (for retexturing, fixing non-manifold vertices);

3. Applying gltf-transform to further optimize the mesh data and apply mesh-optimizer compression OR export in Blender as .glb;

This ensured the models were web-ready and easy to upload and work with further in Semantic Kompakkt. The presentation inspired the Semantic Kompakkt team to work further with the museum team and apply their learnings in the implementation of a post-processing pipeline directly in the Semantic Kompakkt client.

Next, Tim Prins from Studio Stadt presented the Rooted Living research project developed in collaboration with RWTH Aachen and participants around the world to collect 3D scans of different sections of traditional homes, focusing on the transition from the public to private area. Gathering a wide range of architectural building patterns to develop a typology, the project benefits from the Semantic Kompakkt environment in being able to annotate and connect 3D scans with additional contextual information, including integration of AAT concepts, as well as 2D images, such as building blueprints. One of the outcomes of the project were also publications highlighting the results of the research. In this context, the Computational Publishing Service pipeline from NFDI4Culture can also play a role in automating the process of gathering data in a single platform such as Semantic Kompakkt and outputting data dynamically in a variety of publishing formats. A collaboration and testing phase between Rooted Living and CPS is one of the planned outcomes of the workshop.

Last but not least, Sviatoslav Drach and Konstantina Eleftheriadi presented 3D models of seals from the Digital Byzantine Studies (DiBS) project at the University of Cologne and the Cologne Centre for eHumanities. They highlighted their use-cases for working with 3D models in the context of research as well as teaching and learning processes, and they shared their approach to using the XML-TEI / SigiDoc standard for encoding metadata on seals. While they found out that the XML-TEI structure is not directly reusable in the Wikibase Linked Open Data environment of Semantic Kompakkt, they highlighted some useful cases where the Semantic Kompakkt team can work towards better integration of external data standards and data annotations. Furthermore they shared important insights for expanding the existing data model of Semantic Kompakkt specifically concerning inscriptions that may need versioning, translation and more. A final point of discussion that transitioned smoothly into the afternoon discussion about software requirements was the need to provide enhanced data editing capabilities directly in Semantic Kompakkt, to avoid the learning curve of needing also Wikibase and OpenRefine knowledge in order to upload contextual data around the 3D models.

Workshop highlights and common requirements

In the afternoon, event participants had the opportunity to further share their insights from the whole programme and map out a set of common challenges and software requirements. Below we include a short summary of key points gathered during the discussion via a Miro board:

Key learning points:

• The research potential of 3D Models in different fields;
• The possibilities of the RTI technique for different objects – e.g. from Byzantine seals to garment textiles;
• The added value of animations for 3D models;
• The flexibility of metadata properties to meet the needs of different collection projects.

Common challenges:

• Need to further classify images with a typology beyond just 2D or 3D, but also accounting for types of representation – model, blueprint, photo, etc;
• Need for more ways of organizing collections and ways to share filtered views of the 3D models;
• Processing diverse models and large files without built-in post-processing;
• Lack of more informative error messaging and user feedback when errors occur during uploading objects and adding metadata in Semantic Kompakkt. 

Common requirements:

• Iron-out remaining issues within the Kompakkt viewer relating to the upload of models, e.g. perspective selection;
• Develop landing page for institutions and projects featured on Semantic Kompakkt;
• Develop possibilities to include contextual information for researchers on the genesis of the 3D model, e.g. how certain reconstruction decisions were made; 
• Develop possibilities to connect multiple versions of a 3D model to the same physical object; 
• Develop relatable scale and measurements settings;
• Provide use case examples for multi-format publishing of collection catalogues;

Outlook and next steps

In total at the end of the programme, 165 new objects from across the participating projects were uploaded to Semantic Kompakkt by the institutional collaborators with various degrees of assistance from the Semantic Kompakkt team. The entire programme was an iterative learning experience supporting the agile development of Semantic Kompakkt – showcasing areas that need improvement, raising new challenges and providing inspiration for future development. 

An updated roadmap for next steps in the project and what users can expect can be viewed here. Highlights include the development of landing pages for collections and institutions; improved post-processing within the upload workflow; improved support for interoperability between different viewers – including support for the upcoming IIIF 3D specification and a pipeline for data export and exchange between Sketchfab – Kompakt – Europeana. 

For questions or expression of interest in the ongoing work on the project, get in touch with the Semantic Kompakkt team. 
 

Presentation slides:

Dr Sabine de Günther – “Contextualized garments in the digital realm”

Roberta Spano – “Collection of scientific instruments and teaching aids”

Sviatoslav Drach, Konstantina Eleftheriadi – “3D models of byzantine seals”