Luiz do Valle Miranda

Luiz do Valle Miranda, Krzysztof Kutt, Elżbieta Sroka et al.

The rich metadata created nowadays for objects stored in libraries has nowhere to be stored, because core standards, namely MARC 21 and Dublin Core, are not flexible enough. The aim of this paper is to summarize our work-in-progress on tackling this problem in research on cultural heritage objects at the Jagiellonian University (JU). We compared the objects' metadata currently being collected at the JU (with examples of manuscript, placard, and obituary) with five widespread metadata standards used by the cultural heritage community: Dublin Core, EAD, MODS, EDM and Digital Scriptorium. Our preliminary results showed that mapping between them is indeed problematic, but we identified requirements that should be followed in further work on the JU cultural heritage metadata schema in order to achieve maximum interoperability. As we move forward, based on the successive versions of the conceptual model, we will conduct experiments to validate the practical feasibility of these mappings and the degree to which the proposed model will actually enable integration with data in these various metadata formats.

Krzysztof Kutt, Luiz do Valle Miranda, Jakub Gomułka et al.

In response to several cultural heritage initiatives at the Jagiellonian University, we developed a new digitization workflow in collaboration with the Jagiellonian Library (JL). The solution is based on easy-to-access technological solutions -- Microsoft 365 cloud with MS Excel files as metadata acquisition interfaces, Office Script for validation, and MS Sharepoint for storage -- that allows metadata acquisition by domain experts regardless of their experience with information systems. The ultimate goal is to create a knowledge graph that describes the analyzed collections, linked to general knowledge bases, as well as to other cultural heritage collections, so careful attention is paid to the high accuracy of metadata and proper links to external sources. The workflow was evaluated in two pilot studies and in two workshops, which allowed for its refinement and confirmation of its correctness and usability for JL. The knowledge graph created as a result of these pilot studies was made available in a public git repository. As the proposed workflow does not interfere with existing systems or domain guidelines regarding digitization and basic metadata collection in a given institution, but extends them in order to enable rich metadata collection, not previously possible, we believe that it could be of interest to all GLAMs.

Klaudia Ropel, Krzysztof Kutt, Luiz do Valle Miranda et al.

We developed a proof-of-concept method for the automatic analysis of the structure and content of incunabula pages. A custom dataset comprising 500 annotated pages from five different incunabula was created using resources from the Jagiellonian Digital Library. Each page was manually labeled with five predefined classes: Text, Title, Picture, Table, and Handwriting. Additionally, the publicly available DocLayNet dataset was utilized as supplementary training data. To perform object detection, YOLO11n and YOLO11s models were employed and trained using two strategies: a combined dataset (DocLayNet and the custom dataset) and the custom dataset alone. The highest performance (F1 = 0.94) was achieved by the YOLO11n model trained exclusively on the custom data. Optical character recognition was then conducted on regions classified as Text, using both Tesseract and Kraken OCR, with Tesseract demonstrating superior results. Subsequently, image classification was applied to the Picture class using a ResNet18 model, achieving an accuracy of 98.7% across five subclasses: Decorative_letter, Illustration, Other, Stamp, and Wrong_detection. Furthermore, the CLIP model was utilized to generate semantic descriptions of illustrations. The results confirm the potential of machine learning in the analysis of early printed books, while emphasizing the need for further advancements in OCR performance and visual content interpretation.