TITAN

Documents

AIDAR Addressing and Indexing Multimedia Documents Assisted by Voice Recognition Techniques

Frédéric Beaugendre - Voice Insight - October 2006

Project objectives

  • Thematic indexing of “raw” sound streams for radio data
  • Processing without prior data preparation
    • Involves the use of statistical techniques, working on large quantities of data
    • Use of large quantities of labeled data for the learning phase
  • Possible uses
    • Indexing of potentially all types of audio multimedia documents, with possible extension to other languages
    • Radio/Television: (semi-)automatic indexing of archives, assistance with prior art recovery