SuperLectures.com

FAST SPEAKER DIARIZATION BASED ON BINARY KEYS

Full Paper at IEEE Xplore

Speaker Diarization

Přednášející: Xavier Anguera, Autoři: Xavier Anguera, Telefonica I+D, Spain; Jean-François Bonastre, University of Avignon, France

Splitting a speech signal into speakers is the main goal of a speaker diarization system, which has become an important build- ing block in many speech processing algorithms. Current state of the art systems are able to obtain good diarization error rates, but most of them are rather slow, which is a strong handicap in appli- cations that require overall faster than real-time processing. In this paper we present a novel speaker diarization system which is built following a bottom-up agglomerative clustering approach and based on speaker binary keys, recently proposed for speaker modeling. Af- ter initialization, processing is entirely done over binary vectors and using exclusively binary metrics, which makes the system very fast. On tests performed using all conference meetings datasets released for the NIST RT evaluation campaigns we achieve diarization error rates just slightly worse than a classic acoustic-based system while running over 10 times faster.


  Přepis řeči

|

  Slajdy

Zvětšit slajd | Zobrazit všechny slajdy

0:00:17

  1. slajd

0:00:34

  2. slajd

0:01:00

  3. slajd

0:01:16

  4. slajd

0:03:04

  5. slajd

0:03:35

  6. slajd

0:03:44

  7. slajd

0:04:29

  8. slajd

0:05:17

  9. slajd

0:06:35

 10. slajd

0:07:41

 11. slajd

0:07:58

 12. slajd

0:09:30

 13. slajd

0:09:40

 14. slajd

0:10:25

 15. slajd

0:11:14

 16. slajd

0:12:06

 17. slajd

0:12:18

 18. slajd

0:12:31

 19. slajd

0:13:04

 20. slajd

0:13:37

 21. slajd

0:13:52

 22. slajd

0:14:37

 23. slajd

0:15:04

 24. slajd

0:16:35

 25. slajd

0:16:59

 26. slajd

0:17:19

 27. slajd

0:17:57

 28. slajd

0:18:47

    12. slajd

  Komentáře

Please sign in to post your comment!

  Informace o přednášce

Nahráno: 2011-05-24 14:45 - 15:05, Panorama
Přidáno: 15. 6. 2011 15:03
Počet zhlédnutí: 38
Rozlišení videa: 1024x576 px, 512x288 px
Délka videa: 0:19:32
Audio stopa: MP3 [6.60 MB], 0:19:32