A Fuzzy Logic Approach to Handwriting Recognition
D. J. Ostrowski, Pys Cheung · 1996
This paper describes an approach to handwriting recognition using fuzzy logic, and is concerned with cursive script recognition that puts no constraint on the writer. Various other techniques enforce some limitations in order to simplify the recognition task. Commercial experience now shows that although this may be acceptable for notepad type applications, it is essential that the writer should use his or her most natural writing style for any more than brief (about 10 words) notes. The basic styles are illustrated in Figure 10.1. In other recognizers the writer may be constrained to input letters in boxes, making recognition easier, but input slow and tedious. Or writing may be partially or wholly unconstrained, making input fast and natural, but making recognition considerably more difficult.