CFP: Speech Communication Special Issue on Auditory Inspired Spectro-Temporal Features

[Martin Heckmann <Martin.Heckmann@xxxxxxxxxxx>]

Dear AUDITORY list,

this call for papers is for a special
issue of Speech Communication on the topic of Auditory Inspired Spectro-Temporal
Features. The special issue is a follow up of the special session with
the same name to be held on Wed. 24th of September at INTERSPEECH 2008.
We are soliciting contributions not only from participants of the special
session but also from other researchers working in this area. 

Tentative deadline for papers is December
12th 2008

Martin

--------------------------------------------------------------------------------------

Call for Papers: 

Special Issue of Speech Communication

on

Auditory Inspired Spectro Temporal Features

----------------------------------------------------------------------------------------

Aims and Scope  
--------------------      
Inspired by the stunning capacity of
humans to understand speech in the most difficult situations many attempts
have been made to incorporate knowledge from psychoacoustics and psycholinguistics
into computer models. 
With the advances in methodology in
neuroscience we are ready to go one step further. For example, it was recently
possible to measure the receptive fields in the primary auditory cortex
of the cat. Additionally, several studies could identify significant similarities
between the primary visual and auditory cortices. This in turn stipulated
new research as the primary visual cortex is much better investigated than
the primary auditory cortex.

When looking at the ferret it could
be shown that the receptive fields found in the primary auditory cortex
have, as their counterparts in the visual cortex, Gabor like shapes and
respond to modulations in the time-frequency domain. Current features used
for speech analysis, e.g. MFCCs, mainly cover the spectral properties of
the signal. 
It is well known that this lack of temporal
information in the MFCCs, and almost all other features conventionally
used for speech recognition, leads to problems for the recognition of quickly
changing events such as plosives. Moreover, formant transitions, a key
aspect in the perception of speech, are also only covered indirectly by
the MFCCs.

Using spectro-temporal features, similar
to those found in the primary auditory cortex, should overcome the above
mentioned problems. Yet their use leads to new challenges, especially the
selection of the right features from a huge set of features. Nevertheless,
in recent years quite a few people have started to use spectro-temporal
features for tasks such as speech recognition, speech/non-speech discrimination,
and source separation. The results obtained with spectro-temporal features
are promising, especially in noisy conditions. However, they are just a
start and much more research is needed.

The present issue will highlight some
of the neurophysiological underpinnings of such spectro-temporal features
and focus on the different approaches followed so far to make these findings
available to speech processing. This issue is a follow up of the special
session with the same name to be held at INTERSPEECH 2008. 

For this special issue we are soliciting
contributions not only from participants of the special session but also
from other researchers working in this area. 

If you have further questions please
contact me.


Guest Editor
---------------
Martin Heckmann
Honda Research Institute Europe GmbH
Carl-Legien Strasse 30
D-63073 Offenbach am Main
Germany
(martin.heckmann@xxxxxxxxxxx)

Tentative Schedule
----------------------
Papers due 12th December,  2008
Reviews from December 2008 to February
2009
Revisions from March to May 2009
Final papers ready in June 2009

Submission Guidelines
---------------------------
Authors should consult the "Guide
for Authors", available online, at http://www.elsevier.com/locate/specom
for information about the preparation of their manuscripts. Authors, please
submit your paper via http://ees.elsevier.com/specom,
choosing âSpecial issue: Spectro Temporal Featuresâ as the Article Type.
If you are a first time user of the system, please register yourself as
an author.