印鉴纹理识别系统分析
|
摘要
印鉴识别是金融系统工作的一个重要内容，而印鉴纹理识别的识别技术一直是人们广泛关注的关键技术。本文通过对印鉴纹理识别技术、细节特征值等方面的理解提出了一种关于特征值匹配的样条曲线检测方法，从而能够对印鉴纹理曲线作出有效的判断识别。
识别是根据曲线的复杂度及加密方式的不同分别采用广义霍夫变换和傅立叶变换求出频谱等方法建立曲线识别模型。对于期望曲线仅能表述为少量参数的曲线，在把信息光学中有关波形传播的一些特性引入印鉴识别过程后，主要利用在采样阶段中检测到的样本情形，在几何特征和频谱特征等特征相结合的基础上，从曲线相位、幅度、基线、能量谱等识别特征入手，构建一个参数曲线描述，从而部分解决了在较小区域内对印鉴中存在的纹理图像进行配准和特征提取的问题，并对印鉴纹理曲线加以判断识别并得出鉴别结论这一要求。
另外，本论文也证明了因为复杂曲线在较小区域内的较微小的扰动都有可能引起鉴定结果的较大差异，所以贝赛尔曲线不适合于在可识别区域较小(4.5mm~2-10mm~2)时的印鉴识别过程。
It is an important contents of the financial system work that an official seal identify, but an official seal veins identify of identify the technique to has been the key technique that the people pay attention to extensively. The veins detail characteristic withdraws the process to separate in the picture and background, the veins strengthen, picture two the values turn, an official seal picture is thin to turn to examine etc. with the detail characteristic the step constitute.For an official seal veins curve of born, identify is the lines of a certain degree in fact born, examine. This text mainly discussed a curve of kind examination method that concerning the characteristic value match, thus making a judgment to identify to an official seal veins curve.
Identify is at availably born defend the higher curve clan of false and dependable after foundation up, according to complications of the curve and encrypt the way different to adopt the broad sense hough transformation to sign leaf' s transformation to beg a frequency chart respectively etc. the method establishment curve identifies the model, for expect the only ability form of curve to say for the curve of a little amount parameter, according to the sample situation examine in sample stage in the frequency chart characteristic and several the foundation that characteristic combine together up set up a parameter curve description. Thus take into the judgment to an official seal veins curve, get a conclusion with higher accuracy.
Moreover, this thesis also proved because the complicated curve is in smaller district of smaller move to all cause to authenticate the bigger difference of result probably, so can't apply the shell match curve in at can identify the district is smaller of an official seal identify the process.
识别是根据曲线的复杂度及加密方式的不同分别采用广义霍夫变换和傅立叶变换求出频谱等方法建立曲线识别模型。对于期望曲线仅能表述为少量参数的曲线，在把信息光学中有关波形传播的一些特性引入印鉴识别过程后，主要利用在采样阶段中检测到的样本情形，在几何特征和频谱特征等特征相结合的基础上，从曲线相位、幅度、基线、能量谱等识别特征入手，构建一个参数曲线描述，从而部分解决了在较小区域内对印鉴中存在的纹理图像进行配准和特征提取的问题，并对印鉴纹理曲线加以判断识别并得出鉴别结论这一要求。
另外，本论文也证明了因为复杂曲线在较小区域内的较微小的扰动都有可能引起鉴定结果的较大差异，所以贝赛尔曲线不适合于在可识别区域较小(4.5mm~2-10mm~2)时的印鉴识别过程。
It is an important contents of the financial system work that an official seal identify, but an official seal veins identify of identify the technique to has been the key technique that the people pay attention to extensively. The veins detail characteristic withdraws the process to separate in the picture and background, the veins strengthen, picture two the values turn, an official seal picture is thin to turn to examine etc. with the detail characteristic the step constitute.For an official seal veins curve of born, identify is the lines of a certain degree in fact born, examine. This text mainly discussed a curve of kind examination method that concerning the characteristic value match, thus making a judgment to identify to an official seal veins curve.
Identify is at availably born defend the higher curve clan of false and dependable after foundation up, according to complications of the curve and encrypt the way different to adopt the broad sense hough transformation to sign leaf' s transformation to beg a frequency chart respectively etc. the method establishment curve identifies the model, for expect the only ability form of curve to say for the curve of a little amount parameter, according to the sample situation examine in sample stage in the frequency chart characteristic and several the foundation that characteristic combine together up set up a parameter curve description. Thus take into the judgment to an official seal veins curve, get a conclusion with higher accuracy.
Moreover, this thesis also proved because the complicated curve is in smaller district of smaller move to all cause to authenticate the bigger difference of result probably, so can't apply the shell match curve in at can identify the district is smaller of an official seal identify the process.
引文
1、Ueda K, Nakamura Y. Automatic Verification of Seal-Impression Patterns.In:proc of the 7th International Conference on Pattern Recogniton, 1984,2:1019-1021.
2、Fan T J, Tsai W H. Automatic Chinese Seal Identification. Computer Vision, Graphics and Image Processing, 1984,25(3):311-330.
3、Lee S, Kim J H. Unconstrained Seal Imprint Verification Using Attributed Stroke Graph Matching. Pattern Recognition, 1989,22(6):653-644.
4、胡庆，杨静宇，张黔，刘克。基于知识的印鉴鉴别方法。自动化学报，1991，17(6)：696-704
5、张黔，胡庆，杨静宇，蒋韧。统计和结构模式识别方法结合多特征印鉴真伪鉴别方法。计算机学报，1995，18(3)：190-198
6、傅德胜，刘珍丽。印鉴图像配准与特征提取。模式识别与人工智能，2002，15(3)，382-384
7、Nilson NJ. Principle of Artificial Intelligence. Tioga Publishing Co, 1980.
8、Fukunaga K, Introduction to Statistical Pattern Recognition. New York:Academic, 1972.
9、何斌，马天予，王运坚，朱红莲。数字图像处理，北京：人民邮电出版社，2003
10、贾永红。数字图像处理。武汉：武汉大学出版社，2003
11、夏德琛，傅德胜。现代图像处理技术及应用。南京：东南大学出版社，1995
12、孙家广，杨长贵。计算机图形学，北京：清华大学出版社，1995
13、孙家广，杨长贵。计算机图形学，北京：清华大学出版社，1995
14、沈清，汤霖．模式识别导论．长沙：国防科技大学出版社，1991
15、Milan Sonka. Vaclav Hlavac, Roger Boy1:Image Processing, Analysis, and Machine Vision, Second Edition, Thomson Learing and PT Press,1998
16、信号处理原理 中国电力出版社 John G Proakis/dimitris G Manolakis
17、刘宏，李锦涛。基于SUM和纹理的笔迹鉴别，计算机辅助设计和图形学学报，2004，12：1479-1481
18、基于DOG模型的线条检测算法8-9
19、R C Nelson. finding line segments by stick growing[J],IEEE Transactions on Pattern Analysis and Machine Intelligence, 1994, 15(5): 519-523 J-W Lee, I-s Kweon. Extraction of line features in a noisy image[J]. Pattern Recognition, 1997, 30(10): 1651-1660
20、李继云，耿乐丰。给定误差下最少数据点B样条曲线拟合曲线拟合的算法研究，计算机辅助设计与图形学学报．2003，15(4)
21、潘日晶。任意次有理Bezier曲线／面对其控制多边形／网格的整体或局部逼近，应用数学学报。2000，23(2)22、Roger S. Pressman, Software Engineering-A practitioner's Approach, MoGraw-hill inc. 1992
23、邬弘毅。一类连续的B样条曲线及其插值与逼近性，应用数学学报。1999，22(1)
24、刘勇奎、周晓敏。逐点生成参数曲线的双步算法，计算机辅助设计和图形学学报．2002，14(7)
25、邹长春。曲线Hough的变换检测理论。计算机工程与应用。2002，4。
26、Don Chamberlin。Using the new DB2:IBM's object-relational Database System, Morgan Kaufman Publishers inc. 1996
27、张可村，赵英良。计算机数值计算的算法与分析，西安：西安交大研究生数学从书，2003
28、邓贤照、徐英凯。傅立叶变换浅谈。
29、罗纳德N．布拉斯维尔著，杨燕昌译。傅里叶变换及其应用。
30、吴捷，刘延林。图像的定向插值放大及其简化算法。电脑开发与应用，2003，6，22-23
31、郭海涛，田坦，刘国枝，卢逢春。基于模糊聚类的沉底小目标声呐图像的识别方法，小型微型计算机系统，2002，23(8)
32、Iver Jacobson, Object-Oriented Software Engineering-A use Case Diver Kaufman Publishing Company, 1994
33、彩色公文中的公章检测研究。模式识别与人工智能，2000，13(1)，111-113
34、基于模板学习的掌纹识别。计算机应用研究，2003，7，64-65
35、刘胜兰，周儒荣，张丽艳。三角网格模型的特征线提取，计算机辅助设计和图形学学报，2003，15(4)，
36、Lu Yi, TatuLeinonen. Computer simulation of approximate dimensional synthesis with four-bar linkage[J]. Journal of Computer-Aided Desigh&compnter Graphics, 2002, 14(6)
37、李德仁，黄平等．基于Dempster-Shafer特征融合的影响纹理分类．武汉测绘合计大学学报．2000(8)
38、王文惠，孟兵，万建伟等。利用不变量进行基于内容的图像检索。电子学报2002，30(7)945-948
39、周骥、石敦英、赵友兵。图像特征点匹配的强壮算法，计算机辅助设计与图形学学报。2002，14(8)
40、Mirth J A. Quasi presision position synthesis of four-bar linkageds{J}.Transactions of the ASME, 1996,115(1):6571
2、Fan T J, Tsai W H. Automatic Chinese Seal Identification. Computer Vision, Graphics and Image Processing, 1984,25(3):311-330.
3、Lee S, Kim J H. Unconstrained Seal Imprint Verification Using Attributed Stroke Graph Matching. Pattern Recognition, 1989,22(6):653-644.
4、胡庆，杨静宇，张黔，刘克。基于知识的印鉴鉴别方法。自动化学报，1991，17(6)：696-704
5、张黔，胡庆，杨静宇，蒋韧。统计和结构模式识别方法结合多特征印鉴真伪鉴别方法。计算机学报，1995，18(3)：190-198
6、傅德胜，刘珍丽。印鉴图像配准与特征提取。模式识别与人工智能，2002，15(3)，382-384
7、Nilson NJ. Principle of Artificial Intelligence. Tioga Publishing Co, 1980.
8、Fukunaga K, Introduction to Statistical Pattern Recognition. New York:Academic, 1972.
9、何斌，马天予，王运坚，朱红莲。数字图像处理，北京：人民邮电出版社，2003
10、贾永红。数字图像处理。武汉：武汉大学出版社，2003
11、夏德琛，傅德胜。现代图像处理技术及应用。南京：东南大学出版社，1995
12、孙家广，杨长贵。计算机图形学，北京：清华大学出版社，1995
13、孙家广，杨长贵。计算机图形学，北京：清华大学出版社，1995
14、沈清，汤霖．模式识别导论．长沙：国防科技大学出版社，1991
15、Milan Sonka. Vaclav Hlavac, Roger Boy1:Image Processing, Analysis, and Machine Vision, Second Edition, Thomson Learing and PT Press,1998
16、信号处理原理 中国电力出版社 John G Proakis/dimitris G Manolakis
17、刘宏，李锦涛。基于SUM和纹理的笔迹鉴别，计算机辅助设计和图形学学报，2004，12：1479-1481
18、基于DOG模型的线条检测算法8-9
19、R C Nelson. finding line segments by stick growing[J],IEEE Transactions on Pattern Analysis and Machine Intelligence, 1994, 15(5): 519-523 J-W Lee, I-s Kweon. Extraction of line features in a noisy image[J]. Pattern Recognition, 1997, 30(10): 1651-1660
20、李继云，耿乐丰。给定误差下最少数据点B样条曲线拟合曲线拟合的算法研究，计算机辅助设计与图形学学报．2003，15(4)
21、潘日晶。任意次有理Bezier曲线／面对其控制多边形／网格的整体或局部逼近，应用数学学报。2000，23(2)22、Roger S. Pressman, Software Engineering-A practitioner's Approach, MoGraw-hill inc. 1992
23、邬弘毅。一类连续的B样条曲线及其插值与逼近性，应用数学学报。1999，22(1)
24、刘勇奎、周晓敏。逐点生成参数曲线的双步算法，计算机辅助设计和图形学学报．2002，14(7)
25、邹长春。曲线Hough的变换检测理论。计算机工程与应用。2002，4。
26、Don Chamberlin。Using the new DB2:IBM's object-relational Database System, Morgan Kaufman Publishers inc. 1996
27、张可村，赵英良。计算机数值计算的算法与分析，西安：西安交大研究生数学从书，2003
28、邓贤照、徐英凯。傅立叶变换浅谈。
29、罗纳德N．布拉斯维尔著，杨燕昌译。傅里叶变换及其应用。
30、吴捷，刘延林。图像的定向插值放大及其简化算法。电脑开发与应用，2003，6，22-23
31、郭海涛，田坦，刘国枝，卢逢春。基于模糊聚类的沉底小目标声呐图像的识别方法，小型微型计算机系统，2002，23(8)
32、Iver Jacobson, Object-Oriented Software Engineering-A use Case Diver Kaufman Publishing Company, 1994
33、彩色公文中的公章检测研究。模式识别与人工智能，2000，13(1)，111-113
34、基于模板学习的掌纹识别。计算机应用研究，2003，7，64-65
35、刘胜兰，周儒荣，张丽艳。三角网格模型的特征线提取，计算机辅助设计和图形学学报，2003，15(4)，
36、Lu Yi, TatuLeinonen. Computer simulation of approximate dimensional synthesis with four-bar linkage[J]. Journal of Computer-Aided Desigh&compnter Graphics, 2002, 14(6)
37、李德仁，黄平等．基于Dempster-Shafer特征融合的影响纹理分类．武汉测绘合计大学学报．2000(8)
38、王文惠，孟兵，万建伟等。利用不变量进行基于内容的图像检索。电子学报2002，30(7)945-948
39、周骥、石敦英、赵友兵。图像特征点匹配的强壮算法，计算机辅助设计与图形学学报。2002，14(8)
40、Mirth J A. Quasi presision position synthesis of four-bar linkageds{J}.Transactions of the ASME, 1996,115(1):6571