[hts-users:01105] Forward: LIPS 2008: Visual Speech Synthesis Challenge
FYI
---------- Forwarded message ----------
From: Sascha Fagel <sascha.fagel@xxxxxxxxxxxx>
Date: Jan 16, 2008 1:56 PM
Subject: [SynSIG] LIPS 2008: Visual Speech Synthesis Challenge
To: synsig@xxxxxxxxxxx
(apologies for multiple postings)
LIPS 2008: Visual Speech Synthesis Challenge
LIPS 2008 is the first visual speech synthesis challenge. It will be
held as a special session at INTERSPEECH 2008 in Brisbane, Australia
(http://www.interspeech2008.org). The aim of this challenge is to
stimulate discussion about subjective quality assessment of synthesised
visual speech with a view to developing standardised evaluation
procedures.
In association with this challenge a training corpus of audiovisual
speech and accompanying phoneme labels and timings will be provided to
all entrants, who should then train their systems using this data. (As
this is the first year the challenge will run and to promote wider
participation, proposed entrants are free to use a pre-trained model).
Prior to the session a set of test sentences (provided as audio, video
and phonetic labels) must be synthesised on-site in a supervised room. A
series of double-blind subjective tests will then be conducted to
compare each competing system against all others. The overall winner
will be announced and presented with their prize at the closing ceremony
of the conference.
All entrants will submit a 4/6 (TBC) page paper describing their system
to INTERSPEECH indicating that the paper is addressed to the LIPS
special
session. A special edition of the Eurasip Journal on Speech, Audio
and Music
Processing in conjunction with the challenge is also scheduled.
To receive updated information as it becomes available, you can join the
mailing list by visiting
https://mail.icp.inpg.fr/mailman/listinfo/lips_challenge. Further
details will be mailed to you in due course.
Please invite colleagues to join and dispatch this email largely to your
academic and industrial partners. Besides a large participation of
research groups in audiovisual speech synthesis and talking faces we
particularly welcome participation of the computer game industry.
Please confirm your willingness to participate in the challenge, submit
a paper describing your work and join us in Brisbane by sending an email
to sascha.fagel@xxxxxxxxxxxx, b.theobald@xxxxxxxxx,
gerard.bailly@xxxxxxxxxxxxxxxxx
Organising Committee
Sascha Fagel, University of Technology, Berlin - Germany
Barry-John Theobald, University of East Anglia, Norwich - UK
Gerard Bailly, GIPSA-Lab, Dpt. Speech & Cognition, Grenoble - France