臺灣博碩士論文加值系統

隨著科技的日新月異，人們開始使用各種生物特徵來做辨識。利用生物特徵作辨識的應用範圍相當的廣泛，門禁管理、資料控管、機場安全檢驗…等等。生物特徵包含有指紋、掌紋、人臉、虹膜…等等，在這些生物特徵中虹膜具有較多的資訊，並且從人一出生後就定型不再改變，也擁有相當的唯一性。是各種生物特徵辨識技術中相當有用的生物特徵。

此論文提出了一個新穎的方法，使用橢圓近似方程式在YCbCr色彩空間上，搭配瞳孔定位作虹膜切割。在特徵部份，使用小波邊緣偵測基於D4小波轉換，以小波相位振幅(wavelet-phase-amplitude, WPA)作為辨識的依據。以下說明此論文的主要步驟：首先，我門將虹膜影像利用YCbCr色彩轉換，將轉換過的數值投影在三維空間中觀察，根據分佈情形設計一橢圓近似方程式來適合這些分佈，總共切為六個區域，依各個區域而有不同的參數設定。為去除眼瞼部分，我們使用梯度濾波器(Sobel Filter)來修正虹膜的外徑，與瞳孔模板來修正虹膜的內徑。在虹膜辨識部分我們將切割好的虹膜影像經過小波邊緣偵測，取出影像的小波相位振幅紋路，再用對數極座標對應(Log-Polar Mapping, LPM)與能量分析找到主要的紋路分佈，將此紋路所對應到的小波相位及振幅利用二進制作編碼。最後將編出的虹膜碼作為資料庫的測試，我們所使用的測試資料庫為UBIRIS虹膜資料庫，此資料庫總共有241位測試人員1877張虹膜影像，包含有東方人及西方人的虹膜影像。經實驗結果，正確切割率可達96%，正確辨識率為95%。

Along with the changing technology with each day, researchers begin to use all kinds of biometric characteristics to do the identification. It is so expensive that range of applications making the identification using the biometric characteristics, such as the entrance guard management, the personal information security, or the airport security. The biometric characteristics include fingerprint, palm, human face, retina, and iris, where the last one has more effectiveness amidst biometrics because of its uniqueness in the biometric recognition.

The algorithm starts with the proposed elliptic mapping on YCbCr color model and is sequentially accomplished using pupil localization, iris segmentation, wavelet edge characterization via D4 wavelets, and iris recognition by virtue of the novel wavelet-phase-amplitude (WPA) routing. The detail can be described as the following steps. First, we approximate the distribution of the iris image in the YCbCr color space by using our proposed elliptic formula with six parameter settings to fit lighting variations. The eyelashes can be deleted effectively after this manipulation. Sobel filtering is used to locate the outer radius while the inner radius is circularized through the pupil template. To perform iris texture characterization for recognition, the iris subimage is transformed into phase and amplitude images using D4 wavelets, respectively. After log-polar mapping and dominant energy analysis, the representative iris textures are taken to code in four-digit binary format for phase and amplitude images, respectively. Finally, the abovementioned code, short for WPA, is used to generate the link path between phase and amplitude for performing the recognition task on the test set. The UBIRIS iris database is selected to work as our test dataset, which is composed of 1877 images collected from 241 persons including like Asian, American, and European. Experimental results show that the segmentation rate is 96% and the recognition rate is 95% accuracy, respectively.

目 錄

中文摘要 ------------------------------------------------------------ i
英文摘要 ------------------------------------------------------------ iii
誌謝 ------------------------------------------------------------ v
目錄 ------------------------------------------------------------ vi
圖目錄 ------------------------------------------------------------ viii
表目錄 ------------------------------------------------------------ x
一、 序論------------------------------------------------------ 1
1.1 簡介------------------------------------------------------ 1
1.2 研究動機------------------------------------------------- 6
1.2.1 虹膜影像切割之相關研究----------------------------- 7
1.2.2 虹膜影像辨識之相關研究----------------------------- 8
1.3 系統流程------------------------------------------------- 9
1.4 章節概要------------------------------------------------- 10
二、 虹膜前處理---------------------------------------------- 11
2.1 虹膜資料庫---------------------------------------------- 13
2.1.1 UBIRIS虹膜資料庫------------------------------------- 14
2.2 彩色空間模型------------------------------------------- 21
2.2.1 RGB色彩空間------------------------------------------- 22
2.2.2 YUV色彩空間------------------------------------------- 23
2.2.3 YIQ色彩空間-------------------------------------------- 24
2.2.4 YCbCr色彩空間---------------------------------------- 25
2.2.5 HIS色彩空間-------------------------------------------- 27
2.3 虹膜色彩分佈------------------------------------------- 28
2.4 橢圓近似方程式---------------------------------------- 32
2.4.1 橢圓方程式---------------------------------------------- 33
2.5 虹膜內徑與外徑修正----------------------------------- 36
2.5.1 瞳孔中心------------------------------------------------- 37
2.5.2 瞳孔模版------------------------------------------------- 39
2.5.3 梯度分析切割------------------------------------------- 42
2.5.4 虹膜外徑------------------------------------------------- 44
三、 彩色特徵抽取------------------------------------------- 48
3.1 正規化---------------------------------------------------- 50
3.2 小波邊緣偵測------------------------------------------- 56
3.2.1 拉普拉斯濾波器--------------------------------------------- 57
3.2.2 相位及振幅---------------------------------------------- 59
3.3 對數極座標轉換---------------------------------------- 64
3.4 能量分析------------------------------------------------- 70
四、 虹膜辨識------------------------------------------------- 74
五、 實驗結果與討論----------------------------------------- 79
5.1 實驗主架構流程------------------------------------------ 79
5.2 實驗結果與討論---------------------------------------------- 88
六、 結論與未來工作---------------------------------------- 94
參考文獻 ------------------------------------------------------------ 95


圖 目 錄

圖1-1 生物特徵------------------------------------------------------- 1
圖1-2 行為特徵------------------------------------------------------- 2
圖1-3 眼球構造圖---------------------------------------------------- 3
圖1-4 取像裝置示意圖----------------------------------------------- 4
圖1-5 紅外線攝影機擷取虹膜影像-------------------------------- 6
圖1-6 一般相機擷取虹膜影像-------------------------------------- 6
圖1-7 系統流程圖---------------------------------------------------- 10
圖2-1 眼睛的構造---------------------------------------------------- 11
圖2-2 理想的虹膜影像----------------------------------------------- 12
圖2-3 虹膜前處理流程----------------------------------------------- 13
圖2-4 UBIRIS虹膜資料庫------------------------------------------- 16
圖2-5 UBIRIS虹膜資料庫之虹膜顏色分佈-------------------------- 19
圖2-6 使用線性方程式切割虹膜範圍--------------------------------- 20
圖2-7 線性切割虹膜----------------------------------------------------- 21
圖2-8 三原色所混合出來的顏色-------------------------------------- 22
圖2-9 RGB色彩空間---------------------------------------------------- 23
圖2-10 HSI色彩模型圓柱作標系統------------------------------------ 28
圖2-11 膚色於YCbCr與HSV色彩空間比較------------------------------ 31
圖2-12 虹膜顏色於YCbCr中之分佈---------------------------------------------- 32
圖2-13 橢圓近似切割虹膜範圍--------------------------------------------- 32
圖2-14 橢圓修改示意圖----------------------------------------------------- 34
圖2-15 利用近橢圓方程式切割虹膜--------------------------------------- 36
圖2-16 只取出瞳孔的虹膜影像--------------------------------------------- 38
圖2-17 瞳孔部分二值化----------------------------------------------------- 39
圖2-18 瞳孔模板-------------------------------------------------------------- 40
圖2-19 有破損的瞳孔-------------------------------------------------------- 40
圖2-20 利用膨脹侵蝕修補瞳孔--------------------------------------------- 41
圖2-21 兩次膨脹與侵蝕再經過瞳孔模版搜尋後的結果---------------- 41
圖2-22 投影法找瞳孔中心與直徑------------------------------------------ 42
圖2-23 經過橢圓近似切割後的虹膜影像--------------------------------- 43
圖2-24 梯度分析切割說明圖------------------------------------------------ 44
圖2-25 利用梯度濾波並切割的虹膜影像--------------------------------- 45
圖2-26 切割結果-------------------------------------------------------------- 47
圖3-1 特徵抽取流程---------------------------------------------------- 49
圖3-2 極座標轉換示意圖------------------------------------------------ 51
圖3-3 虹膜影像正規化示意圖------------------------------------------ 53
圖3-4 座標轉換示意圖------------------------------------------------- 54
圖3-5 正規化結果-------------------------------------------------------- 55
圖3-6 影像邊緣微分示意圖-------------------------------------------- 57
圖3-7 常用拉普拉斯影像遮罩------------------------------------------ 59
圖3-8 一階微分的矩陣--------------------------------------------------- 61
圖3-9 小波邊緣偵測-------------------------------------------------------- 63
圖3-10 對數極座標對應示意圖--------------------------------------------- 67
圖3-11 對數極座標轉換----------------------------------------------------- 69
圖3-12 3×3模版(Mask)------------------------------------------------ 70
圖3-13 能量分佈圖----------------------------------------------------------- 72
圖3-14 能量分析圖----------------------------------------------------------- 73
圖4-1 虹膜辨識流程圖--------------------------------------------------- 74
圖4-2 相位編碼示意圖----------------------------------------------------- 75
圖4-3 振幅編碼示意圖----------------------------------------------------- 77
圖5-1 主架構流程圖----------------------------------------------------- 81
圖5-2 能量分佈圖----------------------------------------------------------- 86
圖5-3 虹膜辨識系統實際編碼--------------------------------------------- 87
圖5-4 切割輸出影像-------------------------------------------------------- 91
圖5-5 手動切割虹膜影像-------------------------------------------------- 92
圖5-6 虹膜編碼示意圖----------------------------------------------------- 93








表 目 錄

表4-1 相位編碼對應表-------------------------------------------------- 76
表4-2 振幅編碼對應表----------------------------------------------------- 78
表5-1 橢圓近似方程式參數(1)-------------------------------------------- 82
表5-2 橢圓近似方程式參數(2)----------------------------------------- 83
表5-2 正確切割率比較----------------------------------------------------- 84

參考文獻
[1]Anil k. Jain, Arun Ross and Salil Prabhakar, “An introduction to biometric recognition,” IEEE Transaction on Circuits and System for Video Technology, vol. 14, no.1, January 2004.
[2]Institute of Automation, Chinese Academy of Sciences. CASIA iris image database, 2004. http://www.cbsr.ia.ac.cn/Database.htm
[3]Multimedia University. MMU iris image database, 2004. http://pesona.mmu.edu.my/~ccteo/
[4]University of Bath. University of Bath iris image database, 2004. http://bath.ac.uk/elec-eng/research/sipg/irisweb/database.htm
[5]M. Dobes and L. Machala. UPOL iris image database, 2004. http://phoenix.inf.upol.cz/iris/
[6]H. Proenc¸a and Lu´ıs A. Alexandre, “UBIRIS: A noisy iris image database,” 13th International Conference on Image Analysis and Processing (ICIA2005), pp. 970-977, Sep. 2005. http://iris.di.ubi.pt
[7]J. Daugman, “Face and gesture recognition: overview,” IEEE Transactionson Pattern Analysis and Machine Intelligence, vol. 19, no. 7, pp. 675-676, 1997.
[8]J. Daugman and C. Downing, “Epigenetic randomness, complexity and singularity of human iris patterns,” Proc. R. Soc. Lond. B, pp. 1737-1740, 2001.
[9]J. Daugman, “High confidence recognition of persons by iris patterns,” IEEE, pp. 254-263, 2001.
[10]J. Daugman, “How iris recognition works,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 14, no. 1, pp. 21-30, 2004.
[11]R. Wildes, J.C. Asmuth, G.L. Green, S.C. Hsu, R.J. Kolczynski, J.R. Matey and S.E. McBride, “Machine-vision system for iris recognition,” Machine Vision and Applications, vol. 9, pp. 1-8, 1996.
[12]R. Wildes, “Iris recognition: An emerging biometric technology,” Proceedings of the IEEE, vol. 85, no. 9, pp. 1384-1363, Sep. 1997.
[13]C.H. Morimoto, T.T. Santos and A.S. Muniz, “Automatic iris segmentation using active near infra red lighting,” Computer Graphics and Image Processing, 2005. SIBGRAPI 2005. 18th Brazilian Symposium on IEEE, pp. 37-43, Oct. 2005.
[14]GZ. Xu, ZF. Zhang and YD. Ma, “Automatic iris segmentation based on local areas,” Pattern Recognition (ICPR’06), vol. 4, pp. 505-508, 2006.
[15]Asheer K.Bachoo and Jules-R Tapamo, “A Segmentation Method to Improve Iris-based Person Inentification,” IEEE AFRICON Conference, vol. 1, pp. 403-408, 2004.
[16]X. Liu, K.W. Bowyer and P.J. Flynn, “Experiments with an improved iris segmentation algorithm,” Automatic Identification Advanced Technologies IEEE, pp. 118-123, Oct. 2005.
[17]L. Masek and P. Kovesi. “Recognition of Human Iris Patterns for Biometric Identification,” In The School of Computer Science and Software Engineering, The University of Western Australia, 2003. http://www.csse.uwa.edu.au/~pk/studentprojects/libor/.
[18]C.T. Chu and C.H. Chen, “High performance iris recognition based on 1-D circular feature extraction,” IPPR Conference on Computer Vision, Graphics and Image Processing, pp. 73, 2005.
[19]W.S. Chen, P.J. Tsai, S.W. Shih, C.M. Hsieh and S.H. Chang, “An iris recognition technique based on multi-feature extraction and support vector machine,” IPPR Conference on Computer Vision, Graphics and Image Processing, pp. 46, 2005.
[20]L.W. Liam, A. Chekima, L.C. Fan, and J.A. Dargham. “Iris recognition using self-organizing neural network,” Student Conference on Research and Development(SCOReD) Proceedings, pp. 169–172, 2002.
[21]C. Garcia and G. Tziritas, “Face detection using quantized skin color regions merging and wavelet packet analysis,” IEEE Trans. Multimedia, vol. 1, no. 3, pp. 264-277, September 1999.
[22]鐘國亮, “影像處理與電腦視覺,”東華書局, 2004年2月.
[23]CCIR: California Coalition for Immigration Reform (CCIR 601-1). http://www.ccir.net
[24]N. Kazakova, M. Margala and G. Nelson, “Sobel edge detection processor for a real-time volume rendering system,” Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on vol. 2, pp. 913-916, 23-26 May 2004.
[25]E. Boonchieng, W. Boonchieng and R. Kanjanavanit, “Edge-detection and segmentation methods for two-dimensional echocardiograms,” Computers in Cardiology, pp. 541-544, 2004.
[26]J. Chittooru, R. Munasinghe and A. Davari, “Edge detection and segmentation for machine vision”, Proceedings of the Thirty-Seventh Southeastern Symposium on System Theory, pp. 457-461, 2005.
[27]Rafael C. Gonzalez and Richard E. Woods, “Digital image processing,” Prentice Hall, 2002.
[28]S. Mallat and S. Zhong, “Characterization of signals from multiscale edge,” IEEE Transaction on Pattern Analysis and Machine Intelligence, vol. 14, no. 7, pp. 710-732, July 1992.
[29]L. H. Koh, S. Ranganath, and Y. V. Venkatesh, “An integrated automatic face detection and recognition system,” Pattern Recognition, vol. 35, pp. 1259-1273, 2002.
[30]Y. H. Son, B. J. You, S. R. Oh, and G. T. Park, “Affine-invariant pattern using momentums in log-polar images,” International Conference on Automatic face and Gesture Recognition, pp. 797-802, 2004.
[31]W. Xiaohong and Z. Rongchun, “A new method for image normalization,” Proceeding of 2001 International Symposium on Intelligent Multimedia, pp. 356-359, 2001.