with the support of

"Automatic metadata generation allows us to unlock pieces that were never before discoverable."

In the spotlight Geschreven op 30 Jan 2024
Peter

The GIVE-project is a digitisation effort of impressive proportions: over 800.000 newspaper pages, masterpieces and glass plate photographs were digitised and sustainably preserved. With the help of artificial intelligence, over 130 collections with audio and video received more accurate descriptions, making them more searchable. Impressive results, but the project also initiated intensive cooperation between heritage organisations, experts, technical professionals and more.

In this article, we put the spotlight on the people behind the project. This time, we asked our 10 questions to Peter Vanden Berghe, AI architect at Ordina.

Fotocv 1

Can you tell us about your involvement in the GIVE project and your specific role?

I’m the AI architect for the metadata component of the GIVE project. Using artificial intelligence, we were able to add extra descriptions to the heritage objects, making it possible to establish connections across all archives. My role was to support and guide the technical AI team from the research process through the development and production phases of the AI metadata tools. I also had to develop the architecture to implement and integrate the AI metadata pipelines with the existing on-premises and cloud infrastructure.

Why do you think digitising these masterpieces/manuscripts/newspapers/glass plates is so important and valuable?

For me, the metadata aspect of the GIVE project has proven that all the digitisation efforts made in the past have been and will continue to be extremely valuable, even if it wasn’t clear at the time of digitisation what could be done with them. Automatic metadata generation allows us to unlock pieces that were never before discoverable and to use new technologies to gain insights. I only see this increasing in the future.

What impact do you hope the GIVE project will have?

It should give a boost to all other digitisation projects where sceptics question what will be done with the digitised items. Insights may come later with digitisation, but without it, they would probably never come at all.

Sceptics question the value of digitisation, as insights often come later. But without digitisation, those insights would probably never come at all.

What challenges did you encounter in your work on the project, and how did you deal with them?

Processing 170,000 hours of audiovisual content is an overwhelming amount of information. Building AI data pipelines capable of handling this is very challenging, and the problems you intuitively expect often turn out to be less significant than other issues you never anticipated being so troublesome.

What have you learned or discovered during this project?

I’ve learned just how much invaluable material there is in Flanders and how many enthusiastic people are working to preserve and make this heritage accessible.

Which collaborations within the project have made the biggest impression on you, and why?

I was particularly amazed by the collaboration and logistics required to transport and scan the large volume of glass plates.

Metadata

What AI can do for digital archiving

What is your favourite masterpiece, newspaper article, glass plate or piece of metadata, and why?

We had to do a lot of testing on video content to develop our facial recognition pipeline. We often worked with a clip from De Kampioenen for this, and one about Kim Clijsters. I think I’ve watched them hundreds of times now. Another nice clip was the 24-hour theatre performance by Jan Fabre, which was technically challenging because one clip alone lasted 24 hours.

Which masterpiece, manuscript, newspaper or glass plate do you think definitely still needs to be digitised, and why?

I did my PhD research on 3D-printed hip implants, so I have a great affection for anything 3D. I was very charmed by the 3D digitisation of sculptures and hope that – in addition to individual pieces – entire interiors and exteriors of buildings, churches and museums can be digitised.

Who do you think benefits the most or is most impacted by the digitisation of heritage objects?

Primarily, the custodians of the heritage objects and the archivists who work closely with them. But I strongly believe that this digitisation can positively impact wider society. It can engage people with culture and heritage in different ways, and make art more accessible. There’s also a pure technological gain, where we can share our research and development with others, even outside the heritage sector.

Based on your personal experience and expertise, what recommendations would you make to someone who wants to work on a similar project in the future?

It’s very important to involve the different content partners in the development process. We did this through working groups, and the insights we gained there allowed us to see various perspectives and iteratively improve our algorithms.

Read more

Fmrub Br WQAMQ Fc C
In the spotlight

"It’s uncharted territory but incredibly efficient!"

Agoria 2
About the GIVE project

A double award for the GIVE-project

IMG 5158
In the spotlight

"Photographs are always subjective views of reality."