Published on

Published in: Technology, Travel, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • SAMT

    1. 1. Towards dynamic access to audiovisual archives Johan Oomen ~ SAMT2008 ~ Netherlands Institute for Sound and Vision
    2. 2. June 7, 2009 Nederlands Instituut voor Beeld en Geluid
    3. 3. Digitisation as driver for change <ul><li>> User expectations & business opportunities </li></ul><ul><li>> Interoperability ‘beyond retrieval’ </li></ul><ul><li>> Way material can be indexed </li></ul><ul><li>> Position in the production chain </li></ul>Implications for the annotation workflow
    4. 4. Broadcast Professional Public Web Acces Education Medialounge (import) metadata (import) content metadata (conversies) content (encoding) Digital Archive iMMix Digital Facility for Broadcasting YouTube and Open Licences platform Digital TV
    5. 5. The Digital Archive Next 6 years 137.200 hours of video 22.510 hours of film 123.900 hours of audio 2.900.000 photo’s Yearly ingest ~10.000 hours of video ~40.000 hours of radio Digital-born: Dutch television Images for the future Europe’s largest AV digitisation project € 175 million >1,5 petabyte per year
    6. 8. Digital Legacy
    7. 9. The archive as Application Service Provider
    8. 10. Query types <ul><li>A known needle in a known haystack </li></ul><ul><li>A known needle in an unknown haystack </li></ul><ul><li>An unknown needle in an unknown haystack </li></ul><ul><li>Any needle in a haystack </li></ul><ul><li>The sharpest needle in a haystack </li></ul><ul><li>All the needles in a haystack </li></ul><ul><li>Things like needles in any haystack </li></ul><ul><li>Affirmation there are no needles in the haystack </li></ul><ul><li>Most of the sharpest needles in a haystack </li></ul><ul><li>Let me know whenever a new needle turns up </li></ul><ul><li>Where are the haystacks? </li></ul><ul><li>Needles, haystacks, whatever </li></ul>
    9. 11. Need for automatic indexing <ul><li>Economic driver </li></ul><ul><ul><li>Goal: 80 percent automatic / 20 percent manual </li></ul></ul><ul><li>Fine-grained access </li></ul>
    10. 12. Speech to text: Radio Oranje
    11. 13. Visual datamining Ontology of 1000 concepts
    12. 14. <ul><li>Missie Afghanistan uiterst onzeker </li></ul><ul><li>Steeds meer partijen beginnen te twijfelen aan de voorgenomen missie van 1100 Nederlandse soldaten naar Afghanistan. Morgen komen er twee hoge functionarissen van het Pentagon en het State Department naar Den Haag voor overleg met Nederlandse topambtenaren. Vrijdag hakt het kabinet zo goed als zeker de knoop door. Het lijkt een ware worsteling te worden. </li></ul>GTAA Case 1. Semantic annotation GTAA-concept:missie GTAA-concept:militairen GTAA-altlabel:soldaten GTAA-altlabel:kabinetten GTAA-concept:regeringen V é ronique Malaisé Luit Gazendam Hennie Brugman Guus Schreiber Johan Oomen Mettina Veenstra Annemieke de Jong Key contributors CHOICE project
    13. 15. Document Support System
    14. 16. Keyword: Ambassador Location: Uganda Description: This program is about the visit of the US president to the US’s representative of Uganda Query: I’m looking for documents about a visit to a diplomat in Africa GTAA Case 2. Query expansion Diplomat Ambassador Is_a Urganda Africa Is_part of
    15. 17. GTAA Case 3. Query expansion using Wordnet Laura Hollink MuNCH project Research conducted by
    16. 18. Crowdsourcing
    17. 20. Annotation tools Production environment Thesaurus Context documents TV programme Cataloguer Validates Relates to User Generated Metadata Catalogue Description enriches creates
    18. 21. From pilots to production pilots production Sound and Vision Asset Management System sandbox
    19. 22. Video Active: providing Access to TV heritage <ul><li>eContetplus programme </li></ul><ul><li>10.000 video items by 2009 </li></ul><ul><ul><li>Contextual data i.e. stills, programme guides </li></ul></ul><ul><li>10 languages </li></ul><ul><li>HISTORY OF TELEVISION IN EUROPE and EUROPEAN HISTORY ON TELEVISION </li></ul>
    20. 25. Multilingual thesaurus Words entered in simple or advanced search text fields also searches the thesaurus. you can use e.g. the Danish term for animal (“dyr&quot;) and get results in any language, because animal is a thesaurus keyword. Click on keywords to narrow down the search.
    21. 26. Video Active architecture <ul><li>The Video Active backend comprises of various modules, all using web technologies. </li></ul>
    22. 27. `a common access point to Europe's distributed digital cultural heritage` Europeana
    23. 28. 4 million objects 1.000 institutions
    24. 31. Thank you for your attention Johan Oomen |