This paper presents a fuzzy logic approach to efficiently perform unsu
pervised character classification for improvement in robustness, corre
ctness, and speed of a character recognition system. The characters ar
e first split into seven typographical categories. The classification
scheme uses pattern matching to classify the characters in each catego
ry into a set of fuzzy prototypes based on a nonlinear weighted simila
rity function. The fuzzy unsupervised character classification, which
is natural in the representation of prototypes for character matching,
is developed and a weighted fuzzy similarity measure is explored. The
characteristics of the fuzzy model are discussed and used in speeding
up the classification process. After classification, the character re
cognition which is simply applied on a smaller set of the fuzzy protot
ypes, becomes much easier and less time-consuming.