Path: news.mathworks.com!not-for-mail
From: "Minh Tran" <overseer47@gmail.com>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Using wavelet decomposition function dwt2
Date: Sun, 8 Nov 2009 15:50:03 +0000 (UTC)
Organization: The MathWorks, Inc.
Lines: 9
Message-ID: <hd6pbb$84e$1@fred.mathworks.com>
References: <hd63pb$bgc$1@fred.mathworks.com> <hd6ad8$fnm$1@fred.mathworks.com>
Reply-To: "Minh Tran" <overseer47@gmail.com>
NNTP-Posting-Host: webapp-05-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1257695403 8334 172.30.248.35 (8 Nov 2009 15:50:03 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Sun, 8 Nov 2009 15:50:03 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 2082504
Xref: news.mathworks.com comp.soft-sys.matlab:583351


Thanks, Wayne. This is the image I'm working on and the LL output from dwt2:
http://yfrog.com/eldwtproblemp

And this is the results they got on wikimedia :
http://commons.wikimedia.org/wiki/File:Jpeg2000_2-level_wavelet_transform-lichtenstein.png

So apparently something's not right here. I tried grayscale conversion as the guy on wikimedia did instead of using RGB->HSI conversion and it's still the same. The problem shouldn't be the filter - I mean JPEG2000 is different from db4 but shouldn't be that much, as I tried many other filters too (haar, db, bior, etc.). The situation's the same with all other images - whitewashed/overexposed and of course almost useless LL output.

FYI, I'm using the colorspace converter by Pascal Getreuer (http://www.mathworks.com/matlabcentral/fileexchange/7744) to do RGB->HSI conversion. I run the dwt2 on the intensity image immediately after doing the conversion.