Fuzzy graphs to recognize handwritten Arabic characters

Abstract EN

An automatic off-line character recognition system for handwritten cursive characters is presented.

A set of fuzzy constrained character graph models (FCCGMs), which tolerate large variability in writing, are designed.

These models are graphs, with fuzzily labeled arcs which are used as prototypes for the characters.

A set of rules are applied in sequence to match a character’s tree representation to an FCCGM.

To show the applicability of the system, Arabic handwritings of four writers were used in the learning and testing stages.

The system proved to be powerful in tolerance to variable writing and recognition rate.

Index terms: Cursive characters, off-line character recognition, skeletonization, clustering, tree structure, fuzzy numbers, character graph model.