Micro Neural-Controller for Optical Character Recognition

Abstract

This paper presents a Micro-Neural Controller (MNC) for Optical Character Recognition (OCR). The controller consists of both a multilayered feedforward Artificial Neural Network (ANN) and a von Neumann-type microcontroller. The ANN is supervised and trained using back-propagation. The role of the ANN is to perform character recognition while the microcontroller coordinates the data transfer between the network and the user. The results show that the network recognizes the patterns effectively and the microcontroller is able to execute a demonstration program.