Digital library (DL) support for different information seeking strategies (ISS) has not evolved as quickly as the amount of content or the quality of presentation. However, several studies argue for the support of explorative ISS alongside the directed query-response paradigm. Hence, this project presents a primarily explorative research system prototype for metadata harvesting allowing multimodal access to DL content for researchers during the research idea development phase, i.e., while the information need (IN) is vague. To address evolving INs, the prototype also allows ISS transitions, e.g., to OPACs, if accuracy is needed.
As its second contribution, the project presents a curated data set for digital humanities researchers that is automatically enriched with metadata derived by different algorithms including content-based image features. The automatic enrichment of bibliographic metadata is needed to support the exploration of large metadata corpora as traditional metadata does not always address vague INs.
This demo clearly shows that use-case-specific metadata facilitates the interaction with large metadata corpora.
Double-clicking a document lets you inspect a cluster. Clicking a document once makes it the current document.
|←||Back to overview|
|ESC||Remove comparison edges|
|b||Build metadata package|
|c||Compare current document to others|
|d||Display details of current document|
|m||Display map of displayed documents|
- Enriched metadata records (only CSV available with 98 MB before publication)
- Graphs (Publisher-Creator-Spatial-Media)
- Graphs (Generic)
- Visual similarity cluster graphs
- Normalized images
- Local features
- David Zellhöfer