Visual Analytics and Computer Vision Meet Cultural Heritage

Over recent decades, the digitisation of cultural archives has been significantly advanced in many different areas. This includes archival data such as baptismal registers or cadastres, works of visual art, photographs, amateur films, and even entire districts of cities. To make these vast collections navigable to both experts and the broader public, user- friendly interfaces are indispensable. These interfaces need effective data visualisations to help users to explore the data, identify similarities and connections, and draw conclusions. Furthermore, to be able to search the data efficiently, automatic analysis methods are also needed that can capture the complex and diverse content. Our project pursues the vision of investigating interactive applications for two scenarios: On the one hand, the interaction of expert knowledge and automatic, AI-supported content analysis to generate maximum benefit, and on the other hand, the communication of this content to the general public via interactive narratives (storytelling). The cooperative doctoral programme maps the content-related goals in five dissertation projects, which are located in three conceptual levels: Metadata Generation, Data Exploration, and Crafting Data Experiences. Our primary attention is on historical archives, emphasizing photography and film. The research areas are: a) Automatic, content-based analysis of historical film material, b) Automatic determination of used camera and lens types in historical photographs, c) Exploratory analysis of historical image collections with limited or missing metadata, and d) Visual analysis with the aim of investigating the relationships among cultural objects across time and space. The insights gained from the first four dissertation topics will pave the way for e) Innovative storytelling techniques, connecting digital data with physical spaces, aiming for a broader audience reach. The project is interdisciplinary in two respects: Firstly, it bridges the gap between computer science and humanities. Secondly, both visual analytics and computer vision / machine learning are involved within computer science. The participating institutions, FH St. Pölten and TU Vienna, each provide about half of the faculty members of the doctoral programme. This setup ensures each doctoral candidate can engage with both institutions, fostering a holistic academic experience and laying the groundwork for potential future partnerships.

No additional funding sources recorded.