A Combined Fuzzy Set Theoretic and Heuristic Method for Character Recognition

Abstract In this paper the fuzzy set theoretic approach has been combined with a heuristic approach for the recognition of handwritten uppercase english alphabets and numerals. Linguistic description of each character is obtained in terms of the number of horizontal strokes -, vertical strokes |, right slant strokes /, left slant strokes \, A curves ∩, u curves U, c curves C, and D curves ⊃. This is done with the help of membership functions for eaoh of the above features. For the recognition a heuristic approach is used. The recognition procedure was tested by a simulation study and a recognition accuracy better than 92 percent was achieved.