Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!zombie.ncsc.mil!news.duke.edu!convex!cs.utexas.edu!uunet!zib-berlin.de!gs.dfn.de!fauern!rrze.uni-erlangen.de!hub-n.franken.de!ark.franken.de!ralf
From: ralf@ark.franken.de (Ralf W. Stephan)
Subject: Re: wavelets and NN
Message-ID: <1994Sep16.105636.419@ark.franken.de>
Organization: his desk writing an article
Date: Fri, 16 Sep 1994 10:56:36 GMT
References: <35987j$r4n@kodak.rdcs.Kodak.COM>
X-Newsreader: TIN [version 1.2 PL2]
Lines: 18

Craig Martin writes:
> My problem is that I cannot seem to generate good, distinct, feature
> vectors using the wavelet (DAUB4 as shown in numerical recipes).  

The best, state-of-the-art, feature extraction method is Rasta-PLP.
Look into the comp.speech FAQ and the OGI Tools.  The method is
described in

  icsi-ftp.berkeley.edu:pub/techreports/1991/tr-91-069.ps.Z

and a PD implementation with many bells and whistles is in

  icsi-ftp.berkeley.edu:pub/speech


ralf
--
You are in a different maze of little articles, all alike.
