Finding event-related clusters of multimedia objects

From Master Projects
Revision as of 01:02, 3 October 2013 by Laroyo (talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search


About Finding event-related clusters of multimedia objects


Description

One of the goals of the Agora project is to provide users access to museum and archive collections through historical event-driven browsing. However, the collections we are using (from the Rijksmuseum Amsterdam and from the Dutch national audiovisual archive Beeld en Geluid) are too big to peruse casually so we would like to identify interesting topics to create a themed browsing experience.

This master's project is about finding ways to identify representative topics in both the Rijksmuseum and Sound and Vision collections. You would start by analysing the collections, clustering topics within the collections themselves, but also the overlap in topics between the collections and find ways to visualise this intuitively.

Tasks

  • collection data analysis with clustering and ranking algorithms
  • result visualisation

Tools and Data

  • Rijksmuseum collection data + vocabularies
  • Beeld en Geluid collection data + vocabularies
  • Event thesaurus + aligned vocabularies
  • Clustering Tools, Weka
  • Visualisation tools (e.g., Exhibit, Google Chart Tools)

Recommended prior knowledge

It helps to be familiar with the topics covered in the following courses: Ontology engineering , Information Retrieval, Human Computer Interaction

Extra Information

Contact Marieke_van_Erp, Lora_Aroyo, or Guus_Schreiber for more information on this project.