Short:        Find double pictures by id's
Author:       Nagilum
Uploader:     nagilum@chillout.org
Type:         gfx/misc
Requires:     OS3.0,Shell
Distribution: Aminet

What's the use ?
You can use this program to find pictures having the same content, 
even when they have different size.

Why PBM?
This program can be used to create ID's for PBM-pictures.
You may ask yourself, why the the hell PBM and not jpeg, iff or png ?
Well, the first reasond is that PBM has a very simple structure,
and the second is that for all important formats exist command-line
converters to PBM.
This program requires 256 Grayscale pictures in PBM to work,
those pictures start with P5 in the header and can easily created
for example from jpeg: djpeg -grayscale ...
This programm was written in about 4 hours and i know it's not very
clean, but it works (even fast).

How does it work ?
The program calculates the average over the whole picture, then splits
the picture into 64 pieces and calculates the average for each piece,
then depending on if the average from the piece is lower/equal 
or higher to the total average it sets a bit in the id to 0 or 1.
64 pieces create 64 bits, this id will be printed as a 16 digits hex number.
Usally the ID stays the same if the picture is scaled.

You could use a script like the one included (setnote) to put the pictures id
in the filenote, after that you could sort the files by the id and easily
remove double pictures. If you want to use pattern for selecting the pictures
just type:  spat setnote Pictures/#?.jpg  (for example).

DISCLAIMER: THE AUTHOR IS NOT RESPONSIBLE FOR ANY DAMAGE OR LOSS
OF DATA OR WHATEVER CAUSED BY PBMKey. ANYONE USES IT WITHOUT WARRANTY.
.