Proper

Abstract

This paper aims that analysing neural network method in character recognition. A neural network is a processing device, whose propose was expectant by the blueprint and functioning of human brain and their mechanism. The proposed solutions focus on applying Bidirectional Associative Memory and Feature Recognition Neural Network Method for character recognition. The core function of which is to regain in a character stored in memory, when an fractional or noisy version of that character is presented. An associative memory is a warehouse of associated characters that are encoded in various form. In auto-association, an input character is associated with itself and the states of input and output units match. When the warehouse is incited with a given imprecise or partial character, the associated character pair stored in its perfect form is recalled. Character recognition techniques are associated a symbolic identity with the image of the character. This problem of duplication of characters by machines (computers) involves the machine printed characters. There is no idle memory containing data and programmed, but each one neuron is programmed and constantly active.