NIKFI / Projects / SEMANTIC SEARCH FOR VIDEO ARCHIVES

SEMANTIC SEARCH FOR VIDEO ARCHIVES

Deep Learning - Based Solution for Digital Films Semantic Indexing

  • Objects recognition in frames
  • Scenery type recognition
  • Actions recognition in moving frames

Search Engine for Digital Video Content

  • Search by keywords combination.
  • Search by examples
    celebrities/faces
    landmarks/scenery
    previously unknown objects.
  • Search by structure
    emporal relationships
    spatial relationships
    illumination conditions.

Applications

  • Stock movies and TV footage conversion from film to digital
    making digitized content search-able
    easy integration with film conversion systems.
  • Fragmentary search engine for stock archives of digitized film materials
    Professional tool for editors and authors of non-fiction movies and TV studios
    Powerful research tool for historic and social sciences.
  • Public video hosting web portals
    Personal video archives organizing
    Culture hermitage accessibility
    Advertising placement by video context.
  • Video surveillance systems
    Objects lookup
    Suspicious activity detection