Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!clyde.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!apple!julius.cs.uiuc.edu!psuvax1!psuvm!barilvm!bimacs!guedalia
From: guedalia@bimacs.BITNET (David Guedalia)
Newsgroups: comp.ai.neural-nets
Subject: Re: self organising nets in speech recognition
Message-ID: <2643@bimacs.BITNET>
Date: 20 Dec 90 08:27:14 GMT
References: <10793.27653a00@ul.ie>
Reply-To: guedalia@bimacs.UUCP (David Guedalia)
Organization: Bar-Ilan University, Israel.
Lines: 19

In article <10793.27653a00@ul.ie> oreillyj@ul.ie writes:
>
>Could anyone give me any references to work carried out on automatically
>segmenting speech signals using neural nets.  Another area that is also
>of interest is that of feature extraction using neural nets.  These two
>sections will hopefully be part of a speech recognition system which is
>currently using neural nets for phoneme recognition.
>

   I think Professor T. Kohenen has done exactly that.  I am not sure
where it is published.  He used a Fast Fourier Transform to preprocess
the voice signal.  Then fed it into a Kohenen Feature Map.  This
system suceded in dividing the Finish language into 19 (I am not
sure?) phonemes.
  Maybe take a look at;

    Kohenen T., Associative Memory and Self Organization.

 David