摘要
本文分五个部分研究了基于手形和手掌纹理进行身份鉴别的方法。
第一部分首先简要的综述了各种生物特征,然后介绍了手形和掌纹这两个生物特征的优势和方法,以及认证和识别的定义和评价方法。最后介绍了本文的主要工作。
第二部分对前人关于手形和手掌身份鉴别的方法进行简单描述和总结。分为手形和手掌两部分分别阐述不同方法的优缺点。
第三部分研究了基于手形的认证方法,针对手形的特点和现有手形认证方法的不足,本章提出了一种新颖的基于有序手形轮廓点匹配的手形认证方法。首先对预处理后的手形图像进行边界跟踪后提取手指的关键特征点分离手指,运用对有序边缘点进行运算的手指归一化算法进行图像标准化,最后运用新型点匹配算法进行自动鉴别。摆脱了固定栓的束缚,运算量小。通过小样本实验检测,该方法在鲁棒性、准确率、和运算量方面具有良好的综合性能。第四部分研究了手掌鉴别,主要分为三个小节。第一:提出了在线掌纹自动系统定位法,重点研究在线掌纹的定位与分割出感兴趣区域的方法。通过比较以轮廓点为中心的圆盘内掌纹区域及背景所占面积的大小来提取手掌中手指的谷点,分割出感兴趣区域。该方法速度快、鲁棒性好且取得了理想的实验结果。第二:研究了掌纹提取,主要对最明显的主线进行提取,用基于局部灰度均值的方法对图像进行二值化初步提取主线特征点,然后运用统计上回归分析的方法进行后置处理,消除噪声点,并提取精确稳定的细化主线。为掌纹匹配提供了一种方法。第三:提出了一种新的线特征描述和匹配方法。在这个方法中,基于手掌灰度图像的局部二值化得到的二值图作为掌纹的线特征;在匹配阶段提出用相关系数来测量手掌图像间的相似程度。首先对掌纹图像经过预处理技术分割出掌纹区域。预处理后的得到的掌纹图像将只有很少的平移旋转误差。然后把手掌灰度图进行局部二值化得到的图像作为对线特征的描述,它不仅包括线的结构特征,还包括线的粗细程度。在匹配阶段,为提高匹配的精确度采取了以下两个策略。一方面,为克服预处理中旋转不足的问题,我们制定了扩大的训练集。另一方面,设计了基于像素面积的比较算法,该算法有较好的错误容忍度。实验结果表明了本方法的有效性。
第五部分是结论,概括了本文的工作意义和今后的研究方向。
In this dissertation there are five sections to study the methods of identity distinguish technology based on hand shape and palmprint.
Chapter 1 first,summarize a variety of biological characteristics and then introduce the advantages and methods of hand shape and palmprint, as well as the definition and evaluation methods of authentication and identification. Finally introduce the main work in this dissertation.
Chapter 2 describe the methods on hand and palm authentication which were done by predecessors and summarize the advantages and disadvantages of each method.
Chapter 3 work over the authentication based on hade shape.According to the hand shape characteristics and the insufficient of the existing hand shape attestation methods, this chapter proposes a new confirmation and recognition method,then it uses figure normalization algorithm to the ordered boundary points. At last, the system can carry on the automatic distinction by using the new points matching method. This method has gotten rid of the fetter of fixed hitch, and has small operands.Through the small sample experiment, we can see this method has a good overall performance in robustness, accuracy and operand.
Chapter 4 work over the authentication based on palmprint which contains five sections. In the first section, we propose an automatic orientation method used in online palmprint. The thin contour lines are traced out by morphological method after binarization,and then, the valley points are extracted out by comparing the palm’s area with the background’s area in a circle centered at contour lines. Then set up a coordinate system according to the valley points to segment the region of interest. Experiment shows that this method is fast and perfect. In the second section, we work over the principal line extraction. First pick-up the principal lines by auto-adapted local binarization, next it carries on multinomial fitting processing to these points to get the thin principal line and connect the breaking points. The experiment indicated that, these methods can effectively remove the noise and connect breaking points; farther more, the result has small effect error and approaches the natural principal lines of the palmprint,which provides a method for palmprint matching. In the last section, we propose a new method for line characterization and palmprint matching. First, use the preprocessing technique to segment the palmprints, after this preprocessing, translation and rotation of the palmprints remain very little. To get the line features a novel approach to line feature representation and matching is proposed. In this approach, the lines are represented by local binarization based on the palmprint’s gray scale, which contains not only the structural information about the lines, but also the information about the thickness of the different lines.In the matching stage, two strategies are adopted to improve the robustness of matching. The first one is that an enlarged training set is established to overcome large rotations problem caused by imperfect preprocessing. The other one is that a matching algorithm based on pixel-to-area comparison has been designed, which has a better fault tolerant ability for slight translations and rotations. The experiments show that the proposed scheme achieves effective results in terms of authentication rate.
Chapter 5 summarize the significance and future research directions of the work.
引文
[1]邬向前,张大鹏,王宽全.掌纹识别技术[M].北京:科学出版社,2006:1-15,29-38.
[2] Jain A K,Ross A,Prabhakar S.An Introduction to Biometric Recognition.IEEE Transactions on Circuits and Systems for Video Technology,2004,14(1):4-20.
[3]付鹏.手形技术识别综述
[4]孙冬梅,裘正定.生物特征识别技术综述[J].电子学报,2001,29(12A):1744-1748.
[5]顾理.Method of Hand Shape Extrcting in Hand Shape Recognition[J].计算机仿真,2005(22):128-132.
[6] Ajay Kumar,Member,IEEE,and DavidZhang,Senior Membor,IEEE. Peronal Recognition Using Hand Shape and Texture[J]. IEEE Transactions on Image Processing,Vol.15,No.8,August 2006:2454-2461.
[7] Erdem Yoruk ,Ender Konukoglu , Bulent Sankur et al.Shape-Based Hand Recognition[J].IEEE Transactions on Image Processing,Vol.15,No.7,July 2006:1803-1815.
[8] Tadej Savi? and Nikola Pave? i? a.Personal recognition based on an image of the palmar surface of the hand[J],Pattern Recognition, Volume 40, Issue 11, November 2007: 3152-3163.
[9] ZHANG D,YOU J.Online Palmprint Identification[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2003,25(9):1041-1049.
[10]戴青云,余英林.一种基于形态小波的在线掌纹的线特征提取方法[J],计算机学报,2003,26(02):234-239.
[11]李洋,庄庆德,周光辉.基于手掌几何特征的身份识别[J],国外电子测量技术,2005年08期:4-8.
[12] Wai Kin Kong,David Zhang.Palmprint Texture Analysis based on Low-Resolution Images for Personal Authentication[J],IEEE International Conference on Patten Recognition, 2002年:238-505.
[13] Jane You,Wenxin Li,David Zhang.Hierarchical palmprint identification via multiple feature extraction[J]. Pattern Recognition,2002(35):847-859.
[14] Anil K.Jain,Arun Ross Sharath Pankanti.A Prototype Hand and Geometry-based erification System[J].Proc. of 2nd Int’l Conference on Audio- andVideo-based Biometric Person Authentication (AVBPA), Washington D.C., March 22-24, 1999:166-171.
[15] Sanchez-Reillo,R.,Sanchez-Avila C., Gonzalez-Marcos. Biometric Identification through Hand Geometry Measurements.IEEE Transactions on Volume 22, Issue 10, Oct 2000 Page(s): 1168–1171.
[16] Kuo-Chin Fan,Chih-Lung Lin,Win-Long Lee. Deformable Matching of Hand Shapes for Verification.16th IPPR Conference on Computer Vision, Graphics and Image Processing,2003.8:17-19.
[17] Li W,Zhang D,Xu Z.Palmprint identification by Fourier transform [J],Journal of Software,2002,16:417-432.
[18] Lu G,Zhang D,Xu Z.Characterization of palmprint Recognition Using Eigenpalms Features[J]. Pattern Recognition Letters, 2003, 24(9-10): 1473-1477.
[19] Duta N,Jain A K,Mardia K.Matching of palmprint [J].Pattern Recognition Letters,2001,23 (4):477-485.
[20] Kong W K,Zhang D,Li Wen-xin.Palmprint feature extraction using 2-D Gabor filters[J].Pattern Recognition,2003(36):2339-2347.
[21]顾理,庄镇泉,万淑超等.手形识别中的手形提取方法[J].计算机仿真, 2005.7第22卷第7期,:128-132.
[22]冈萨雷斯.数字图像处理[M].第二版.北京:电子工业出版社,2003.3:51-105.
[23]飞思科技产品研发中心编著.MATLAB 6.5辅助图像处理[M].北京:电子工业出版社,2003:32-104, 189-209.
[24]郭振滨,裘正定.基于曲线拟合的手形生物特征认证新算法[J].Journal of Computer Research and Development 42(11) 2005:1870-1875.
[25] Zang D,Shu W. Two novel characteristics in palmprint verification:datum point invariance and line feature matching. Pattern Recognition.1999,32:691-702.
[26] Li W,Zhang D,Xu Z. Image alignment based on invariant features for palmprint identification. Signal Processing:Image Communication.2003,18:373-379.
[27] Han Chinchuan,Cheng Hsuliang,Lin Chihlung,et al. Personal authentication using palm-print features[J]. Pattern Recognition, 2003, 36:371-381.
[28]戴青云,余英林,张大鹏.掌纹身份识别系统中的定位分割技术[J].广东工业大学学报2002,19(1):1-6.
[29]盛丽,吕英华,孔俊.小波模极大值的手掌自动识别系统定位法[J].吉林大学学报(信息科学版)2006,5(24):548-554.
[30]戴青云,余英林.一种基于形态小波的在线掌纹的线特征提取方法[J],计算机学报,2003,26(02):234-239.
[31] Wu Xiangqian,Wang Kuanquan,Zhang David. An approach to line feature representation and matching for palmprint Recognition[J]. Journal of Software, 2004, 15(6):869-880.
[32]宋炯,林喜荣,包桂秋等.一种掌纹线特征的提取方法[J].计算机工程与应用, 2004, 12:32-33,57.
