This paper is about recognition of hand written character using artificial neural network. Kohonen self-organization technique is implemented for pattern recognition. By the end of this paper we get conclusion that hand written character recognition is one of the efficient methods of recognition.

According to character recognition it takes input as a character and compares it with predefine character class. The starting step of this research is to convert the written text to computer understandable form; Research has been done with artificial neural networks.

The advantage with the Kohonen self-organization map is that the system is changing according to the changing condition and inputs. Even the recognition ratio is good in the proposed system.

According to the pattern recognition, recognition is based on concrete and abstract items which include two assumptions where class membership identifies input patterns that share common features and other assumption where network identifies input patterns for common features.

As a part of research patterns are taken from handwritten characters with a set of patterns it was found that it was 95% accurate. The main problem with recognition of character is increased with noisy data with various differences in handwriting because of writer or can be due to nature of writing.

In this paper an experiment was conducted on characters which are handwritten. The experiment was carried with 10 different persons and the tests are conducted on the system with combination of characters. The overall accuracy was around 95% with disconnected character.

This paper gives a recognition system based on hand written character. The performance increases by this implementation to fullest. People times cannot identify their own writing; it may depend on various factors so finding efficiency is difficult. And the written text is converted into computer readable form.

Download  Hand Written Character Recognition System Using Kohonen Self Organization Map Paper.