Home |  ENGLISH |  Kontakt |  Impressum |  Anmelden |  KIT

Vortrag3420

Aus Aifbportal

Wechseln zu: Navigation, Suche
Everything you always wanted to Find * but never dared to Search


Informationen zum Vortrag

Datum: 30. April 2017
Titel: Everything you always wanted to Find * but never dared to Search
Referent: Harald Sack

Veranstaltung:
Name: Metadata Management in Film Archive Workshop
Ort: Potsdam
Land: Germany

Forschungsgruppe: Information Service Engineering



Link to the presentation slides

Abstract: Popular Video on Demand platforms are all facing a similar problem: how to provide the users with potentially interesting movies that fit to the users' (current) interests and preferences. Moreover, how enable the users also to achieve an overview of all the available archive content without to overstrain or being boring. Guiding the users through the archive along (previously unknown) paths reflecting the user's interests and putting the presented documents in a content-related perspective cannot be obtained by traditional retrieval, but also requires guided browsing and intelligent recommendations, leading to a so-called exploratory search.

To enable exploratory search in video archives, document content has to be annotated with descriptive and machine understandable metadata. Since manual human effort is limited and costly, automated state-of-the art video and audio analysis methods are applied to obtain descriptive metadata of heterogeneous quality, accuracy, as well as reliability and confidence. For successful metadata integration, semantic analysis relates the (raw) media analysis metadata to knowledge bases and encodes the annotation as reusable and machine understandable data that can be exploited for semantic search, content-based recommender systems, as well as exploratory search.

This presentation will provide a brief overview on state-of-the art technologies of video analysis and subsequent semantic analysis for the automated generation of enriched Linked Data based metadata. Furthermore examples are presented for semantic search systems, content-based recommender systems, and exploratory search scenarios.