O Desafio de Vesúvio é promovido pela Universidade de Kentucky e tem como objetivo incentivar cientistas a desenvolverem algoritmos para escanear pergaminhos em papiro e transformá-los em imagens em alta resolução por meio de tomografia computadorizada e Inteligência Artificial.
O pergaminhos do desafio foram produzidos antes do ano 79, quando erupção do Vesúvio devastou a cidade de Pompéia e os incendiou:
Fonte: https://www.nature.com/articles/d41586-024-00346-8
Usando tomografia computadorizada, os pesquisadores conseguiram fazer o “desenrolamento virtual” dos papiros e, com ajuda de Inteligência Artificial, identificaram partes do texto:
Fonte: https://scrollprize.org/
“Papyrus scrolls can contain more than 200 columns of text, so we have to bear in mind that what we know about this text will be hugely increased by the next results of virtual unwrapping. That being said, we have the lucky opportunity to read the final columns of the scroll, where the author would often summarize the content of his work or introduce new topics that he would address in subsequent books of the same work,” Federica Nicolardi, assistant professor in papyrology at the Università degli Studi di Napoli Federico II, continued. (https://www.engr.uky.edu/news/2024/02/vesuvius-challenge-grand-prize-winners-announced)
Fonte:https://www.engr.uky.edu/news/2024/02/vesuvius-challenge-grand-prize-winners-announced
O desafio e todo o trabalho se desenrola no contexto do EduceLab, iniciativa de restauração digital:
What is EduceLab?
EduceLab is a highly specialized heritage science laboratory expertly designed to provide data-intensive yet object-centric solutions to the most challenging problems in the study of cultural heritage. Its unique ecosystem of non-destructive instrumentation offers key scientific capabilities that are crucial to addressing the challenging variability of heritage science contexts. These capabilities include:
- Materials characterization
- Advanced multimodal imaging (tomography, photography, photogrammetry) with gold standard bench equipment as well as a flexible, configurable prototype environment
- Cyberinfrastructure and methodologies for capturing, structuring, processing, and mining large-scale data sets
- Mobile and flexibly-deployed instrumentation for in-situ data acquisition and on-site evaluations
EduceLab comprises four operational clusters – BENCH, MOBILE, FLEX, and CYBER. Each one is based on usage patterns that match the needs of diverse heritage science communities.
Fontes: