A Design Approach for Hand Written Character Recognition Using Adaptive Resonance Theory Network I

Adaptive Resonance Theory Network I (ART1) is a neural network concerning unsupervised learning. It is the first member of the ART family. ART1 can learn and recognize binary patterns. The basic idea in ART1 is that the input vector is compared to the prototype vectors in order of decreasing similarity until a prototype vector close enough to the input vector is found. In this paper, we are going to recognize the Hand Written Character. The process of recognition is divided into three steps. First the Written Character is pre-processed, then ART1 algorithm is employed to the pre-processed character to extract the Features. In the final stage, the accuracy of ART1 Network is evaluated. This paper shows that ART1 is going to recognize the Character with good accuracy rate.