Handwritten Character Recognition System Using Neural Network

• Main Author: Kuryati, Kipl
• Format: Final Year Project Report
• Language: English
• Published: Universiti Malaysia Sarawak, (UNIMAS) 2004
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://ir.unimas.my/id/eprint/47535/1/Kuryati.pdf; http://ir.unimas.my/id/eprint/47535/

Character recognition has been an area of research for a long period of time. It has been argued that this problem is difficult to be modelled using classical modeling techniques, and that neural network offer a new perspective to approach this problem. Therefore the intention of this project is to investigate the application of neural networks to the problem of recognizing handwritten alphabet and digit characters. In this project, a software system capable of recognizing alphabet and digit characters incorporated with neural network algorithm was developed using MATLAB neural network toolbox. This project also outlines the experimental evidence that have been compiled while investigating possible approaches to character recognition. In addition, performing recognition simulations compare the performances of the various neural networks and the best neural network performance is then chosen. At the end of the project, the most suitable backpropagation network properties setting for character recognition were presented and discussed