Directories:

00-original-images/

	- original images supplied by customer
	  (probably fetched from the website)
	
01-images-with-names/

	- the same images as in the previous case, however their names
	  are altered to their content; for example file with name
	  01-images-with-names/G22B.jpg contains letters "G22B"

02-extracted-fonts/

	- here are stored extracted font files

	- this directory is EMPTY by default, you need to create its contents:

	$ php fonts-extract.php 01-images-with-names 02-extracted-fonts

	- note that it is good idea to erase output directory before extraction

03-output-PHP-code/

	- here is stored output PHP file, it's code contains font definitions
	  stored in the PHP associative arrays

	- this directory is EMPTY by default, you need to create its contents:

	$ php fonts-get-PHP-code.php 02-extracted-fonts 03-output-PHP-code/data.php

And now is everything ready to run OCR test suite. Simply try ./do-ocr.sh,
it will loop through all the image files, compare OCRed contents with their
filename and report if they are the same or differs.

OCR itself is implemented in do-ocr.php file.

Nepto, 2005-04-02