Datasheet

2 - 54
Selecting the proper document language is imperative. Based on the selection
of a language, the software knows which symbol set to recognize. Multilinguis-
tic support ensures that “exotic” characters such as ç, ß, ñ, γ and ø are recog-
nized correctly.
Secondly, the software extensively uses linguistic databases to validate its
results. Suppose that you have to read the word "president" where an ink stain
makes the "r" look like an "f". Looking things up in the English lexicon, the IRISPen
will detect autonomously that the word "president" is being read and that it doesn’t
make any sense to recognize the symbol "f". This “self-learning” technique is
of course highly dependant on the linguistic context.
Linguistics also offer useful help to solve ambiguous cases such as an "O"
which might be mistaken for a '0'. Another typical example is the letter "l" and the
number '1' which have an identical form in many fonts - think of texts produced
on old typewriters! The linguistic context helps to determine whether you are
dealing with "l" or '1'.
3chapter2.p65 2/16/2001, 3:23 PM54