A question (somewhat linux related, but not all the way.)

Jeff Goeke-Smith jeff@goeke.net
Wed, 2 Aug 2000 13:07:54 -0400


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

This is a thread (Somewhat edited for readablility) that I just had with
Daniel Bennett.  I think I have found the solution in a program called
HaruspeX which can be found at http://www.linux.it/ospiti/haruspex/

Thanks for everybody's help.

- --Jeff

- -----Original Message-----
From: Jeff Goeke-Smith [mailto:jeff@goeke.net]
Sent: Wednesday, August 02, 2000 12:02 PM
To: linux-user@egr.msu.edu
Subject: A question (somewhat linux related, but not all the way.)


- -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Ok,  I'm working on a project where I have a large number of images.  Some
of these are repeats of the same image, just scaled or compressed with jpg.
I would like to keep the highest quality of each image, and get rid of the
extras, but as it stands right now, the file names leave me no clue as to
which ones are duplicates.
So, here's the question.  Does anybody know of a software package to do
image comparisons looking for very similar images that might be different
pixel sizes but when scaled are the same image?  Or better yet, does anybody

know of a library that has the necessary functions to pull this off?
Perhaps the GIMP can do it?
If anybody has any ideas.  They are much appreciated.


Thanks,
Jeff

From: Bennett, Daniel [mailto:daniel.bennett@jnli.com]
Sent: Wednesday, August 02, 2000 12:27 PM
To: 'Jeff Goeke-Smith'
Subject: RE: A question (somewhat linux related, but not all the way.)


        Just out of curiosity, what are you doing with all the images?

From: Jeff Goeke-Smith [mailto:jeff@goeke.net]
Sent: Wednesday, August 02, 2000 12:39 PM
To: Bennett, Daniel
Subject: RE: A question (somewhat linux related, but not all the way.)


Long term storage in a relational database.  What it boils down to is that I
have a ton of images that have been taken, scanned, cropped, resized, and
all sorts of other things done to.  I want to load all of them into a large
database and index them based on content.  However, I'm pretty sure about
half of the images are repeats and were modified off the orignal in some
respect.  So It only makes sence to me to drop those before I do the
indexing.  I hate, hate, hate, hate, hate  repetitive work, so I just
thought this is the perfect job for software.

- --Jeff

From: Bennett, Daniel [mailto:daniel.bennett@jnli.com]
Sent: Wednesday, August 02, 2000 12:46 PM
To: 'Jeff Goeke-Smith'
Subject: RE: A question (somewhat linux related, but not all the way.)


    Ever heard of HaruspeX?  It's a GIMP/PostgreSQL plugin that provides
cataloging and searching for similar images.. Beyond that, I've got a lot of
optical physicist/astronomer friends that do stuff with Fourier analysis..
They might have something the performs this sort of task.


From: Jeff Goeke-Smith [mailto:jeff@goeke.net]
Sent: Wednesday, August 02, 2000 12:58 PM
To: Bennett, Daniel
Subject: RE: A question (somewhat linux related, but not all the way.)


THANKS!!!!  Good lord I would have been searching for a while to find that.
Concidering that This is what my final goal actually is, This is great.  It
looks an awful lot like this will do what I want with just a little bit of
hacking.

Again.  Thanks alot.

(Mind if I post this thread back at  the list so people know about it?)

From: Bennett, Daniel [mailto:daniel.bennett@jnli.com]
Sent: Wednesday, August 02, 2000 1:01 PM
To: 'Jeff Goeke-Smith'
Subject: RE: A question (somewhat linux related, but not all the way.)


You're welcome.. Go ahead.

-----BEGIN PGP SIGNATURE-----
Version: PGP for Personal Privacy 5.5.3
Comment: Use PGP, it makes Big Brother wonder what you're up too!

iQA/AwUBOYhVamxe5pKMpy4HEQJDVQCgtT9wggaY0fzw+eJEFiQYqkwXYQ4AoLbp
6r1jTz8l/L/cbMu1/OBTK29p
=G22Q
-----END PGP SIGNATURE-----