Application of Neural Network in Recognition of Hand Written Hangeul
Jeong-Young SONG, Hee-Hyol LEE, Sang-Gu LEE, Kageo AKIZUKI
In this paper, the authors propose a recognition method of hand written Hangeul. Hangeul characters are basically classified to 6 patterns by the positions of their elements. Using these characteristics of Hangeul, the pattern of the given character is determined by its peripheral distribution and the other features, and the character is resolved into its elements which are the vowels and the consonants. They are recognized by the optimized neural network. To optimize the neural network, we discuss optimization of the neural network parameters, such as the inclination of the sigmoid function, the numbers of the input layer's units and the hidden layer's units.
The constructed recognition system is applied to non-learning Hangeul written by some Korean peoples, and recognition rate of 97.6 % is obtained.