  This is a demo of my OCR program. This program is a combined
character recognition and translation program, designed to
translate a graphics '*.TIF' file into a text file.
The main idea behind the recognition scheme is the centre
of gravity technique. Where the centre of gravity of the
pattern to be recognized is found, then the center of gravity of
each of the 8 diagonals is calculated. The result is 18 numbers
identifying the pattern exactly. This pattern is checked
back with the library, if it is found then the character
corresponding to this pattern is outputed to the text file
else it is replaced by a hash, indicating that the program
failed to recognize this character. If in learn mode then
the unrecognized pattern is displyed to the user and the
corresponding character is requested. This new pattern is
learned and stored in the library for future checks.
