Thesis title: Un ecosistema digitale integrato per lo studio delle decorazioni musive bizantine
This research addresses Byzantine geometric mosaic decorations, aiming to bring coherence to a rich yet heterogeneous corpus. Rooted in Architectural Drawing and in dialogue with art history, heritage studies, and computer science, the goal is to define a digital integrated ecosystem that turns visual documentation into structured, queryable, and reusable knowledge. The contribution unfolds along two complementary axes: on the one hand, the automatic classification of motifs, to streamline readings and reduce manual workload; on the other, the relational modelling of outcomes within an atlas/knowledge graph that makes explicit the connections between images and metadata.
The state of the art highlights the fragmentation of lexicons and standards, the predominance of traditional approaches, the limited application of AI techniques, and the absence of integrated, interoperable protocols. In response, the thesis proposes a unified workflow that links acquisition, observation, processing, representation, and visualization, ensuring traceability and critical oversight.
The methodology integrates data curation protocols, AI-assisted recognition procedures, and interpretability tools. Results indicate that, when paired with transparent checks, automation supports coherent and comparable readings across contexts, bringing to light regularities, variants, and borderline cases. In conclusion, the ecosystem proves scalable and reproducible: it opens to new sites and decorative families, fosters shared protocols, and integrates collaborative platforms, with the ambition of making representation an interoperable locus of knowledge for the history of ornament.