Extending an atomistic Fedora-Commons object model to facilitate image segmentation and enhance discovery

We have hundreds of thousands of scanned pages from books, manuscripts, letters, theses, newspapers, scrapbooks, etc. How can we further describe this material logically, while optimizing it for research discovery? This presentation aims to provide a stack-agnostic strategy for extracting new Fedora objects from existing content, and illustrate how to incorporate this new material into existing discovery layer schemes.

Topics include:

- Organize pages from books logically, into chapters
- Create new objects out of articles extracted from newspapers, snippets from scrapbooks, and other items from complex image data
- Find specific image data objects, resources, and hierarchical folders through the discovery layer, without cluttering the query results