ISCA - International Speech
Communication Association


ISCApad Archive  »  2017  »  ISCApad #229  »  Academic and Industry Notes  »  ASVspoof 2017 CHALLENGE: Audio replay detection for automatic speaker verification anti-spoofing

ISCApad #229

Monday, July 10, 2017 by Chris Wellekens

4-10 ASVspoof 2017 CHALLENGE: Audio replay detection for automatic speaker verification anti-spoofing
  
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*

  ASVspoof 2017 CHALLENGE:
  Audio replay detection for automatic speaker verification anti-spoofing

  http://www.spoofingchallenge.org/

=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*

  Are you good at machine learning for audio signals? Are you good at
  discriminating 'fake' signals from authentic ones? Are you looking for new
  audio processing challenges? Do you work in the domain of speaker recognition?
  ASVspoof 2017 challenge might be for you!

CHALLENGE TASK:

  Given a short clip of speech audio, determine whether it contains
  a GENUINE human voice (live recording), or a REPLAY recording (fake).

  You will be provided a development set containing genuine/replay labeled audio
  examples, along with further metadata such as speech content and the devices
  used in the replay recordings. Your task is to develop a system that assigns a
  single 'liveness' or 'genuineness' score value to new audio samples, and to
  execute that system on a set of test files for which the ground truth is not
  provided. We provide a Matlab-based reference baseline method to kick-off
  quickly towards developing your new ideas!

  For more details, refer to the evaluation plan in the website:
  http://www.spoofingchallenge.org/

BACKGROUND:

  The goal of the challenge series is to enhance security of automatic speaker
  verification (ASV) systems from being intentionally circumvented using fake
  recordings, also known as 'spoofing attacks' or 'representation attacks' in
  the context of biometrics. ASVspoof 2017 is a second edition of a challenge
  kicked off in 2015, and the new perspective in ASVspoof 2017 are the replay
  attacks, especially 'unseen' attacks - for instance, containing replay
  environments, devices and speakers that might be very different from those in
  the development data.

  Despite 'ASV' being in the challenge title, you do NOT require knowledge
  of automatic speaker verification: the task is a 'standalone' replay audio
  detection task that can be addressed as a generic acoustic pattern classification
  problem. We welcome as many new ideas to the problem as possible!

SCHEDULE:

  Development data published:   December 23rd, 2016
  Evaluation data published:    February 10, 2017
  Evaluation set scores due:    February 24, 2017
  Results available:            March 3, 2017
  Interspeech paper deadline:   March 14, 2017
  Metadata/keys published:      May 2017
  Interspeech special session:  August 2017

REGISTRATION:

  Send a free-worded e-mail to asvspoof2017@cs.uef.fi
  to register and obtain the dev data.

ORGANIZERS:

  Tomi Kinnunen, University of Eastern Finland, FINLAND
  Nicholas Evans, Eurecom, FRANCE
  Junichi Yamagishi, University of Edinburgh, UK
  Kong Aik Lee, Institute for Infocomm Research, SINGAPORE
  Md Sahidullah, University of Eastern Finland, FINLAND
  Massimiliano Todisco, Eurecom, FRANCE
  Hector Delgado, Eurecom, FRANCE

Back  Top


 Organisation  Events   Membership   Help 
 > Board  > Interspeech  > Join - renew  > Sitemap
 > Legal documents  > Workshops  > Membership directory  > Contact
 > Logos      > FAQ
       > Privacy policy

© Copyright 2024 - ISCA International Speech Communication Association - All right reserved.

Powered by ISCA