Asia-Pacific Youth Conference on Communication Technology (APYCCT 2010 E-BOOK)

Kunming,China,8.7-8.8,2010

ISBN: 978-1-935068-20-4 Scientific Research Publishing, USA

E-Book 934pp Pub. Date: August 2010

Category: Computer Science & Communications

Price: $120

Title: Handwritten Digit Recognition Based on a New Combination Feature
Source: Asia-Pacific Youth Conference on Communication Technology (APYCCT 2010 E-BOOK)(Part 6 Network Theory and Technology) (pp 826-829)
Author(s): Xinbo Zhang, College of Information &Electronic Engineering, Zhejiang Gongshang University, Hangzhou, China
Lili Wu, College of Information &Electronic Engineering, Zhejiang Gongshang University, Hangzhou, China
Abstract: One kind of nine-segment projection features is proposed, the nine-segment projection features belong to statistical feature, they can make the recognition system simple and convenient. The point features belong to structural feature, they can visually reflect characters’ property and characteristic, they have their separate advantages and disadvantages. And the nine-segment projection features and the point features are put into BP neural network separately, finally the handwriting digits can be recognized. The experimental results show all the recognition rates with these single features are lower. According to their advantages and disadvantages, these two kinds of features are combined, after the new combination feature was put into BP neural network, the recognition rate raises obviously. The improved BP algorithm was applied to recognize the handwriting digits, not only the number of iterations reduce , but also the recognition rate with the single features and the combination feature raise further.
Free SCIRP Newsletters
Copyright © 2006-2024 Scientific Research Publishing Inc. All Rights Reserved.
Top