基于模式识别的手写汉字识别毕业论文.doc

基于模式识别的手写汉字识别毕业论文

摘要
模式识别就是通过计算机用数学技术方法来研究模式的自动处理和识别。随着计算机技术的发展，人类对模式识别技术提出了更高的要求。在我国如何将汉字方便、快速地输入到计算机中已成为关系到计算机技术普及的关键问题。特别是对于大量己有的印刷资料和手稿，计算机自动识别输入己成为必须研究的课题，所以汉字识别在文献检索、办公自动化、邮政系统、银行票据处理等方面有着广阔的应用前景。
对手写汉字进行识别，首先将汉字图像进行处理，抽取主要表达特征并将特征与汉字的代码存储在计算机中，这一过程叫做“训练”。识别过程就是将输入的汉字图像经处理后与计算机中的所有字进行比较，找出最相近的字就是识别结果，这一过程叫做“匹配”。
本文主要介绍了模式识别的基本原理和手写的10个汉字字符的识别系统的设计实现过程。第一章介绍了与模式识别有关的概念、描述方法、有关的数学运算和此学科的发展状况。第二章分析了模式识别的过程、图像文件的格式，并详细介绍了BMP格式图像的结构；由于本文设计的识别系统是在VC++环境下开发的，所以给出了CDib类的定义和功能介绍。第三章主要讨论了分类器设计过程中的特征空间优化设计问题、分类器设计准则、基本方法和判别函数的知识。第四章在前三章知识的基础上，给出了模板匹配法识别图像的原理，类GetFeature的功能及定义，识别系统设计的开发流程及实现过程，并对实验结果做出简单的分析。
关键词：模式识别  BMP位图  分类器 手写汉字识别 模板匹配

I
ABSTRACT
The pattern recognition researches how to treat with and recognize pattern automatically through computer with math arithmetic. Along with the development of computer technology, human need more advanced pattern recognition technology. In our country it is a key problem for computer technology’s popularization that how to input Chinese characters into computer fleetly and conveniently. Especially for large numbers of printed data and manuscript, the automatic recognition and input of Chinese characters becomes a stringent task, therefore the Chinese character recognition will have a broad application prospect on literature retrieva l, office automation, postal service system, bank bill processing.
In order to recognize Chinese characters, the first task we have to do is feature extraction of a map, after that we have to store the feature in the computer. This process is called "the training". The second process is called “match”, This process compares the hand-written Chinese character’s feature and the stored features in the computer.
This paper mainly introduces the basic principles of pattern recognition and the design、 realization process of ten hand-written Chinese characters recognition system. The first chapter mainly introduces the concepts related to the pattern recognition, the description method, the mathematics operation and this discipline’s development condition. The second chapter introduces the pattern recognition’s process, the picture format, especially the BMP form which is used in this paper. This article designs the recognition system under the VC IDE, therefore we introduce a CDib class’s definition and the functions. The third chapter mainly introduces characteristic space optimization design, the classification’s design criterion, the essential method and the knowledge about knowledge function. The fourth chapter produces the principle of template match recognition arithmetic, the class  GetFeature’s definition and the functions, the recognition system’s design flow and the realization’s process, finally we produce the experimental result and the simple analysis.
Kewords: Pattern Rocognition BMP  Classification  Template Match
Hand-written Chinese Characters Recognition

II

目 录
第1章 绪 论 1
1.1模式识别中的一些基本概念 1
1.2模式的描述方法 1
1.3模式识别学科介绍及现状 3
1.4论文的主要研究工作 3
第2章 模式识别系统及位图基础 4
2.1模式识别系统结构 4
2.2图像识别 5
2.2.1图像识别简介 5
2.2.2图像识别过程 5
2.3位图基础知识 6
2.3.1图像文件格式介绍 6
2.3.2 BMP文件结构 8
2.3.3 CDib类的建立 10
第3章 分类器设计 12
3.1 特征空间优化设计问题 12
3.2分类器设计准则 13
3.3类器设计基本方法 14
3.4 判别函数 14
3.5训练与学习 16
第4章 模板匹配分类器识别汉字系统的实现 17
4.1系统的功能描述 17
4.2模板匹配法的理论基础 17
4.3特征类的设计 18
4.4系统的设计流程及功能的具体实现 19
4.5实验结果和分析 22
结束语 24
致谢词 25
参考文献 26
附录 27