学科门类: 军事学 单位代码:90009
中图分类号:TP391 密 级:公 开
硕士研究生学位论文
基于数据场的笔迹鉴别研究
一级学科: 军 队 指 挥 学
学科专业: 军 事 通 信 学
研究方向: 指挥自动化理论与技术
培养院所: 计算机与指挥自动化学院
总参第六十一研究所
研 究 生 : 陈 罡
指导老师: 李 德 毅 院士
中国人民解放军理工大学
二 OO 三 年三月
Handwriting Identification Based on
Data Field
A THESIS SUBMITTED TO
PLA UNIVERSITY OF SCIENCE AND
TECHNOLOGY
FOR THE DEGREE OF MASTER
Author: Gang Chen
Supervisor : Prof. Deyi Li
Beijing, March 2003
中国人民解放军理工大学硕士学位论文
第
i
页
摘 要
关键词
中国人民解放军理工大学硕士学位论文
第
ii
页
ABSTRACT
Handwriting identification is a complex and difficult problem that is
important for surveillance and security, telecommunications, digital libraries,
video meeting, and human-computer intelligent interactions. Only the experts
can recognize the identities of people from their handwriting. We are still
puzzled with the psychological and physiological nature of the process.
Cloud model and data field are the theory basis and basic methods to
solve the problem in the thesis. Cloud model is a model of uncertain transition
between a linguistic term of a qualitative concept and its numerical
representation. Data field is the field produced by data, just as electric field is
produced by electric charges. The contour view of the data field produced by
all data in a set by the means of potential function shows clustering apparently.
We apply data field method to the on-line HSV and get four kind of
fields associated with the pressure, pressure changing frequency, velocity, and
acceleration respectively. Then features can be gotten from these four fields
and these features use not only the static location, but also the dynamic
properties. So, preliminary experiments were done by using Chinese
handwriting signatures and very promising results were attained.
By writer identification (WI), the writer of the handwritten document
can be detected by analyzed his handwriting style without regard to content.
We transform the handwriting sample into four characteristics which are
pressure, pressure changing frequency, velocity, and acceleration. We can get
four potential curves from the data field of these characteristics. Then we
extract eigenvectors to distinguish handwritings by these potential curves. The
method is simple and direct. It is independent of content and needs no any
pretreatment.
Keywords: Cloud Model, Data Field, HSV Handwriting
Identification
Gang Chen
Directed by Prof. Deyi Li
中国人民解放军理工大学硕士学位论文
第
iii
页
目 录
第 1 章 绪 论 ............................................................................................... 1
1.1 .......................................................................... 1
1.2 ...................................................................................... 3
1.3 .................................................................. 4
1.4 .......................................................................... 5
1.5 ...................................................................................... 7
1.6 ...................................................................... 8
第 2 章 定性定量转换云模型 ....................................................................... 10
2.1 ........................................................................................................ 10
2.2 .................................................................................................... 10
2.2.1 .................................................................................. 10
2.2.2 .......................................................................................11
2.2.3 .................................................................. 13
2.3 ................................................................................................ 15
2.3.1 .................................................................................. 15
2.3.2 ............................................................................. 16
2.3.3 ............................................................. 17
2.4 ................................................................................................ 18
第 3 章 数据场思想 ..................................................................................... 19
3.1 ........................................................................................ 19
3.1.1 ................................................................................. 19
3.1.2 .................................................................................. 20
3.2 ................................................................................ 21
3.3 σ ................................................................................. 23
3.4 MERCER .................................... 25
3.5 ................................................................................................ 25
第四章 云模型与数据场的关系 ................................................................. 26
4.1 ................................................................................ 26
4.2 ........................................................ 27
4.2.1 .................................................................. 27
4.2.2 .................................................................. 28
4.2.3 .......................................... 29