Optical computing and investigations on writing

In previous years several methods have been developed for extracting significant information parameters contained in written, typed, or printed texts. They have shown that some models (optical, mathematical, or hybrid) could be described for linear and nonlinear operations. At first, essential features of information contained in a written page are quantitatively specified and related to legibility criteria. Then an approach to handwriting analysis is made through coherent optical processing along two distinct ways: (1) averaging permits assessing a degree of resemblance between any letter in a set and that taken as a reference, (2) as similarities cannot be extracted directly, the introduction of a form function applies to the comparison of printed letters of various types taken as a whole and other kinds of graphic signs. Finally a nonlinear model built up from polynomials can afford a complete analysis of handwriting as illustrated by various examples.