摘要
在现代社会中,随着计算机及网络技术的高速发展,信息安全显示出前所未有的重要性。身份鉴定一般可分为三类:基于特定物品、基于特定知识和基于生物特征。前两类方法存在携带不便、容易遗失、或者由于使用过多或不当而损坏、不可读和密码易被破解等诸多问题。因此,目前广泛使用的依靠证件、个人识别号码、口令等传统方法来确认个人身份的技术面临着严峻的挑战,并越来越不适应现代科技的发展和社会的进步。人们希望有一种更加方便可靠的办法来进行身份鉴定。生物特征识别技术给这一切带来可能。生物测定学是通过利用个体特有的生理和行为特征来进行身份识别和个体验证的一门科学。
所有基于生物特征的自动识别系统都有大体相同的工作原理和工作过程。首先是采集样本,这些样本可以是人脸的图像,或者是声音的数字化描述、或者是指纹等;其次是进行特征提取,根据样本所具有的独特和唯一的特征,用一种算法为其分配一个特征代码,这一代码被存入数据库;最后当需要对某人进行身份鉴定时,再用某种特征匹配算法将存入数据库的此人的特征代码与被识别人的特征相匹配,从而查明其身份。
在众多的生物特征识别方法中,虹膜识别具有唯一性、稳定性、防伪性和非侵犯性等特点。本文的虹膜识别算法主要包括以下几个步骤:虹膜图像的分割、归一化、特征提取和编码匹配。
虹膜图像的预处理主要实现虹膜内外边缘的确定,瞳孔的定位,虹膜部分图像的归一化和增强。根据虹膜图像的灰度直方图分割出瞳孔,得到了虹膜的内边缘。然后使用Canny算子提取图像边缘,根据已有的虹膜内边缘使用Hough变换得到虹膜的外边缘。再根据内外圆圆心的位置关系把得到的圆环图像归一化为矩形图像,最后将得到的图像进行直方图均衡化,增加了图像的对比度,消除了光照不均对虹膜图像的影响。把已得到的矩形虹膜图像分块,使用多通道二维Gabor滤波器提取虹膜的形状和纹理结构。根据系数对图像编码。把得到的编码存入编码数据库以备查询。编码匹配基于一种加权的Hamming距离,相同虹膜和不同虹膜的加权Hamming距离分别聚集在不同的值附近,适当的选择阈值就可以区分不同的虹膜。
根据此算法实现了相应的软件,并使用现有的虹膜图像数据库进行试验,识别率可达95%以上,验证了本文算法的有效性。
With the rapid development of new computer and network technologies, information security is becoming more and more important than ever. There are two traditional methods for identification. The first is using certain information such an passwords and PIN numbers. The second is using certain thing such as a key. The two methods are unreliable, because the key may be lost and the password may be forgotten. Biometrics is a kind of science of using individual personal characteristics to verify identity. There are some mature and widespread biometric identification technologies, such as: fingerprint identification, Facial Detection, iris identification, retina identification, palm identification, voice identification, signature identification.
All of the biometric identification technologies have almost the same recognition process. The first work is sample collection. The sample could be images of faces, or records of voices, or fingerprints. Then extract the unique feature from the samples, and calculate codes from a specific algorithm. The codes will be stored in a database.
If there is a new sample need to be identified, we lookup the database to match the code.
Irises possess distinct features for uniquely identifying a person. Iris-Identification is regarded as a kind of noninvasive human identification technique. This paper elucidates in detail the principles and the typical structure of Iris-Identification system. The system includes four parts: segmentation, normalization, feature extraction and recognition.
Image preprocessing begins with precisely locating its inner and outer boundaries. We find the threshold form the histogram which can separate the pupil from the image. After that we perform edge detection algorithm based on the Canny operator, then select the points in the edge to perform Point Hough Transform to make recognize the outer boundary. Then center deviation of iris boundaries is normalized. At last we make histogram equalization to approach contrast enhancement.
By using discrete 2D Gabor filters, we get relative coefficients from the filtered images. Then the pattern matching algorithm based on weight Hamming distance is discussed for iris code.
Experiment results for our iris database prove that the proposed approach for iris recognition is effective. We have got the accuracy rate more than 95%.
引文
1 O’Gorman. Comparing Passwords, Tokens, and Biometrics for User Authentication. Proceedings of the IEEE. 2003, VOL 91:2019~2020
2 Anil K. Jain, Arun Ross, Salil Prabhakar. An introduction to biometric recognition. IEEE Trans on Circuits and Systems for Video Technology. 2004, 14 (1) : 4~20
3 K. Jain and S. Prabhakar and Lin Hong. A multichannel approach to fingerprint classification . IEEE Trans on PAMI. 1999, 21(4): 348-359
4 Blanz V, Vetter T. Face Recognition Based on Fitting a 3D Morphable Model. IEEE Trans on Pattern Analysis and Machine Intelligence. 2003,25(9):1063-1074
5 Plamondon R, Srihari S N. Online and off-line handwriting recognition: A comprehensive survey. IEEE Trans on Pattern Analysis and Machine Intelligence. 2000, 11(1): 68-89
6 Little J, Boyd J. Recognizing people by their gait :The shape of motion. Journal of Computer Vision Research. 1998,1(2):2-32.
7 Furui S. Recent Advances in Speaker Recognition. Pattern Recognition Letters, 1997, 18(9): 859-872
8束为,荣钢,边肇祺等.利用掌纹进行身份自动鉴别方法的研究.清华大学学报(自然科学版).1999,15(1):95-97
9 John Daugman. High Confidence Personal Identification by Rapid Video Analysis of Iris Texture. International Carnahan Conference on Security Technology, Taiwan, China, 1992: 1~11
10汤萍,潘永称,刘继萍. 1761眼屈光结构的生物测量.中国实用眼科杂志. 1995,13(5):289-292
11 Kronfeld P C. The Gross Anatomy and The Embryology of the Eye. The Eye. 1968,1:1-66
12 Richard P. Wildes. Iris recognition: An Emerging Biometric Technology Proceeding fo IEEE. 1997, 85: 1347- 1363
13 John Daugman. Neural Image Processing Strategies Applied in Real-time Pattern Recognition. Real Time Imaging . 1997 , 3 (3) : 157～171
14 John Daugman. Statistical Richness of Visual Phase Information: Update on Recognizing Persons by Iris Pattern. International Journal of Computer Vision. 2001,45(1).25~38
15 John Daugman. Complete Discrete 2-D Gabor Transform by Neural Networks for Image Analysis and Compression. IEEE Trans on ASSP. 1988, 36(7): 1169~1179
16 R. P. Wildes, J. Asmuth, S. Hsu, et al. Automated, Noninvasive Iris Recognition System and Method. US Patent. No. 5572596, 1996
17 W. W. Boles, B. Boashash. A Human Identification Technique Using Imagesof-the Iris and Wavelet Transform. IEEE Trans on Signal Processing. 1998, 46(4):1185~1188
18李连贵,廖曙铮等.一种序列虹膜型图像的质量评价方法.测控技术. 2001.5:27~28
19何家峰,叶虎年,叶妙元.虹膜图像质量评价的研究.中国图象图形学报. 2003,8: 387~391
20王蕴红,朱勇,谭铁牛.基于虹膜识别的身份鉴别.自动化学报. 2002,1:1~10
21叶学义,庄镇泉,李军,张云超.实用虹膜识别算法.生物识别研究新进展.清华大学出版社, 2002:150-156
22苑玮琦,马军防,狄文彬.基于主动轮廓线的虹膜定位方法.计算机工程与应用. 2003,34:104~107
23 N Ostu. A Threshold Selection Method from Gray Level Histograms. IEEE Trans on SMC-9.1979:62~66
24吴冰,秦志远.自动确定图像二值化最佳阈值的新方法.测绘学院学报. 2001,12:283~286
25徐国旺,廖明潮.拟合圆的几种方法.武汉工业学院学报. 2002,4:104~1051
26林金龙,石青云.用点Hough变换实现园检测的方法.计算机工程2003, 7:17~19
27 John Daugman. High confidence visual recognition of persons by a test of statistical independence. IEEE Transactions. 1993,15:1148~1161
28宫雅卓,谭南虹,施鹏飞.虹膜纹理归一化的圆心偏差校正算法.上海交通大学学报. 2004, 4:578~582
29 John Daugman. Recognizing persons by their pattern. Information Security Technical Report. 1998,13:33~39
30 Bresenham J. Algorithm for computer control of digital plotter . IBM System J, 1965,4(1): 25~30
31 Field D. Incremental linear interpolation . ACM Transactions on Graphics. 1985,4: 1 ~11
32肖斌等.一种基于高斯函数的直方图规定化算法.铁道学报. 2006, 8 :119~122
33 Yasumasa Itoh, Tanaka Yutaka. Image enhancement based on estimation of high resolution component using wavelet transform. Proc IEEE International Conference on Image Processing. Paris, 1999:489~493
34 CHENG Heng-da, XU Hui-juan. A novel fuzzy logic approach to contrast enhancement. Pattern Recognition. 2000,33(5):80~819.
35 Stark J. A Adaptive Image Contrast Enhancement Using Generalization of Histogram Equalization . IEEE Trans Image Processing. 2000,9(5): 889~896
36 Zhu H, Chan FHY, Lam F K. Image Contrast Enhancement by Constrained Local Histogram Equalization. Computer Vision and Image Understanding, 1999,73(2):281~290
37田启川,潘泉等.不完全虹膜模式唯一性实验研究.计算机应用研究. 2006,2: 237~239
38 Dennis Gabor. Theory of Communication. J. Inst. Elect. Eng. 1946,93:429~457
39 Dunn D, Higgins W E, Wakeley J. Texture segmentation using 2D Gabor elementary functions. IEEE Transactions on PAMI.1994,16(2):130~149
40 John Daugman. How Iris Recognition Works. The Computer Laboratory, Cambridge CB2 3QG, U.K.:London ,1998
41赵英男,杨静宇.一种改进的Gabor滤波器特征抽取算法及应用.系统仿真学报. 2005, 9: 2236~2259
42 Lim T R, Guntoro A T Thiang. Car recognition using gabor filter feature extraction. Circuits and System. 2002,2:451~455
43 P Kaizinga, N Petkov, Grigorescu. Comparison of texture features based on gabor filters. Proceedings of the 10th International Conference on Image Analysis and Processing. 1999:142~147
44 Zehang Sun, Bebis G, Miller R. On-road vehicle detection using gabor filters and support vector machines. Digital Signal Processing. 2002,2:1019~1021
