AI4DH 2024_3rd International Workshop on Artificial Intelligence for Digital Humanities

Silvia Cascianelli and Giuseppe Cartella participated in the 3rd International Workshop on Artificial Intelligence for Digital Humanities (AI4DH).

Silvia Cascianelli presented the paper “µgat: Improving Single-Page Document Parsing by Providing Multi-Page Context”, co-authored with Fabio Quattrini, Carmine Zaccagnino, Laura Righi and Rita Cucchiara. The work focuses on regesta as visually rich, multi-page documents and extends recent document parsing architectures by incorporating information from neighbouring pages, in order to obtain more accurate, machine-readable representations of complex historical sources.

Giuseppe Cartella presented the paper “Pixels of Faith: Exploiting Visual Saliency to Detect Religious Image Manipulation”, co-authored with Vittorio Cuculo, Marcella Cornia, Marco Papasidero, Federico Ruozzi and Rita Cucchiara. The study addresses the challenge of detecting partially manipulated religious images using deep learning models enhanced by human-derived visual saliency maps and highlights a human-in-the-loop approach to improve fake detection systems, contributing to the preservation of religious and cultural content in the age of generative AI.

Both contributions are closely related to the research activities of the ITSERR project, which promotes the use of artificial intelligence for the analysis, structuring and safeguarding of textual and visual heritage.