Character recognition using digitally implementable neural network

Pao Yue-kong Library Electronic Theses Database

Character recognition using digitally implementable neural network

 

Author: Lee, Kin-sang Timothy
Title: Character recognition using digitally implementable neural network
Degree: M.Sc.
Year: 1996
Subject: Optical character recognition devices
Pattern perception
Neural networks (Computer science)
Hong Kong Polytechnic University -- Dissertations
Department: Multi-disciplinary Studies
Dept. of Electronic Engineering
Pages: iv, 86, 3 leaves : ill. ; 30 cm
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b1555440
URI: http://theses.lib.polyu.edu.hk/handle/200/1223
Abstract: The aim of this project is to investigate the use of digitally implementable neural networks in character recognition. This work begins with a study on the effect of using different types of activation functions for the multi-layer feed-forward neural network. The back-propagation learning algorithm is used for training all the networks that have been tested in this project. Apart from the conventional approach for character recognition that is used by neural network, an alternative approach based on function approximation have been studied and tested, and the performance is found to be satisfactory. The STBP and MS models suggested by Wang [5, 6] have been studied and tested. A 'modified' MS model is proposed to resolve the stability problem of the STBP and MS models. It is found that the 'modified' MS network can be trained to achieve the required accuracy very much faster than those conventional BP networks, and can solve the stability problem anticipated by the STBP and MS models.

Files in this item

Files Size Format
b15554405.pdf 2.984Mb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

     

Quick Search

Browse

More Information