to the VALID Database Website!
THE VALID DATABASE IS NO LONGER AVAILABLE
The VALID Database will be released
to the public at the Audio Visual based Biometric Person Authentication
in July 2005 
Acedemics regiser here
to get a free copy!
performance of deployed audio, face, and multi-modal person
recognition systems in non-controlled scenarios, is typically
lower than systems developed in highly controlled environments.
With the aim to facilitate the development of robust audio,
face, and multi-modal person recognition systems, the large
and realistic multi-modal (audio-visual) VALID database was
acquired in a noisy “real world” office scenario
with no control on illumination or acoustic noise
database consists of five recording sessions of 106 subjects
over a period of one month. One session is recorded in a studio
with controlled lighting and no background noise, the other
4 sessions are recorded in office type scenarios.
identification experiments using visual speech features extracted
from the mouth region were carried out. These are aimed to
be a baseline comparison for similar experiments. The performance
of the uncontrolled VALID database is compared with that of
the controlled XM2VTS database. Example lip region identification
accuracies for VALID and XM2VTS are 63.21% and 97.17% respectively.
The results highlight the degrading effect of an uncontrolled
illumination environment and the importance of this database
in providing real world deployment metrics.
This work is part of the VALID biometric project,
funded by Enterprise Ireland. Andrea Osl and Rosalyn Moran
are thanked for the collection/preparation of the VALID database.
Finally, a special thank you is given to all the database
"The Realistic Multi-modal VALID database and Visual
Speaker Identification Comparison Experiments," N. A.
Fox, B. A. O'Mullane, and R. B. Reilly, to appear in the Proc.
of the 5th International Conference on Audio- and Video-Based
Biometric Person Authentication (AVBPA-2005), New York, July
20-22, 2005. PDF Paper Link