BSC’s Saint George on a Bike AI Art Project Launches Citizen Science Campaign

BSC's Saint George on a Bike AI Art Project Launches Citizen Science Campaign

May 5, 2022 – Analyzing the millions of works of art that are part of the cultural and artistic heritage is a seemingly impossible task for humans, but not so for supercomputers. European project St. George on the bike, coordinated by Barcelona Supercomputing Center (BSC) In cooperation with european foundation, started in 2019 with the aim of using the MareNostrum 4 supercomputer to train artificial intelligence models, which help to spread the wealth of European cultural heritage among citizens and recognize its value as well as understand, recognize and promote its preservation. The project aims to generate automatic descriptions of hundreds of thousands of images from different cultural heritage repositories using natural language processing and deep learning Algorithms.

In the second phase of the project, researchers launched a crowdsourcing campaign In Zooniverse, a citizen science portal based on open peer-to-peer collaboration, to collect thousands of manual annotations that help better train these AI models. The campaign is completely open and anyone can participate by accessing it Link.

Our project will allow rapid access to rich cultural information, which can serve equally well for cultural and social purposes, education, tourism, and possibly for historians or anthropologists. Citizens can indirectly benefit from better public services, when these are based on insight that richer metadata we produce – such as web accessibility for the visually impaired or narratives that can reveal social injustice or inclusion and gender issues through heritage institutions cultural and helping to create a more tolerant European identity.” says Maria-Cristina Marinescu, BSC researcher and Saint George on a Bike project coordinator.

So far, no artificial intelligence system has been built and trained to help describe images of cultural heritage with a maximum coverage of themes, objects and iconographic relationships taking into account the time period and scene formation rules for sacred icons XIV-XIV. eighteenth century.

says Joachim Morey, BSC Researcher and Project Expert in Computational Linguistics. “For example, when you first identify a motorcycle in a fifteenth-century painting of St. George, it corrects itself and identifies the most believable thing of the period, which is the horse. This adaptation will also be made according to the cultural context. For example, in context In Japanese culture, what we call in Europe a knight is a samurai.

The project also launched an inspiring video (below) highlighting the use of artificial intelligence to discover never-before-seen images and installations, extract relationships between thousands of images or the opportunity to curate virtual exhibitions with relevant paintings from around the world.

More information about the project:

Source: BSC

Source link

Leave a Comment