本論文提出基於空間群集模型並使用進階多解析度區塊局部二元化圖形(Advanced Multi-resolution Block Local Binary Pattern)作為特徵點的人臉辨識系統，並以ORL face database以及Extended Yale B face database來進行實驗。
傳統上我們在做人臉辨識多半使用取特徵的方式是以所謂以模型(Model)為基礎的方式來對影像降維並擷取特徵，這類型方式對輸入影像資料的處理多半是所謂全域的(Holistic)，例如：主成分分析(Principle Component Analysis)、線性識別分析(Linear Discriminant Analysis)，會使用這類型的模型來對資料做模擬除了此類型方法早已廣在統計領域上被用於資料分類模擬外，也是因為這類型的方式算是會利用資料目前的相互特性來做模擬，是一種子空間學習的演算法(Sub-Space Learning Algorithm)，可對於後續作分類有加強的效果，但由於此類全域模擬的方法，會造成辨識效果容易受到取像變異的影響，例如：光源或是影像正規化不足等等。所以近年來以往用於圖型識別的紋理描述器(Descriptor)漸漸受到關注應用於人臉辨識或是其他生物辨識上，例如：賈伯濾波器(Gabor Filter)、局部二元化圖形(Local Binary Pattern)。
會使用紋理描述器來取得我們需要的影像特徵原因是因為這類型的紋理描述器所取得影像的特徵資訊是以局部的方式取得(Local)，這就代表著我們可以避開一些以全域模擬影像資料的缺點，例如：部分的影像資訊破壞(Local Occlusion)以及影像正規化不足的問題，且此類的描述器描述出來的資訊受光源的影響較不是那麼的敏感。本篇論文使用局部二元化圖形(Local Binary Pattern)為基礎，改進傳統局部二元化圖形的一些缺點，進而利用這些不同局部二元化圖形擷取出來特徵點在影像上分布的情形來做為我們辨識的依據。我們利用可以良好模擬自然界現象的高斯混合化模型(Gaussian Mixture Model)來模擬取得這些特徵點分布的資訊。運用高斯混合化模型取得的參數，我們建構出可以良好抵抗位移(Shift)、尺寸(Scale)、旋轉(Rotation)的特徵資訊，最後我們運用多類支持向量機(Multi-Class Support Vector Machines)做訓練產生分類器，配合投票的方式決定辨識的結果以及對非系統的人做過濾。

This research presents a comprehensive recognition system based on spatial constellation model using advanced multi-resolution block local binary patterns. We perform object-oriented design code to build our system and do the experiments of face recognition using ORL and Extended Yale B face database.
We usually do dimension reduction and get features from facial images based on model-based method for the conventional face recognition. This kind of methods is holistic when they deal with the image data, for example, PCA (Principle Component Analysis), LDA (Linear Discriminent Analysis) and so on. Researcher use this kind of methods is because they are good methods to do dimension reduction in statistics. They find the relation of input data which are good for we to represent the data. However, these holistic model methods are easily affected by the variance of images we get such like illumination change and insufficient normalization. In the past few years, some descriptors like Gabor Filter and local binary pattern used in pattern recognition are noticed and be used in face recognition or other biometrics. Because these descriptors get the features from images are local, that is why we use it. We can avoid insufficient normalization problem that usually happen in holistic method. Our research is based on LBP (local binary pattern) and we use Gaussian Mixture Model, it’s a good method to model the phenomena in the nature world, to model the distribution of the characteristic points in spatial domain. Using parameters of GMM we get, we can construct good features which can robust against the distortion of input image like variation of shift, scale, and rotation. We use multi-class
II
support vector machines to train our data in the database, and we make the final decision by counting the votes of each support vector machine.

ABSTRACT I
CONTENTS III
LIST OF FIGURES V
LIST OF TABELS VIII
CHAPTER1 INTRODUCTION 1
1.1 THE PURPOSE AND BACKGROUND OF RESEARCH 1
1.2 MOTIVATION OF RESEARCH 2
1.3 ARCHITECTURE OF PAPER 3
SECTION 2 FACE RECOGNITION 4
2.1 FACE DETECTION 4
2.2 FACE FEATURE EXTRACTION 5
2.3 CLASSIFICATION 20
SECTION 3 PROCESS OF RESEARCH AND ALGORITHM 28
3.1 ARCHITECTURE OF FACE RECOGNITION SYSTEM 28
3.2 RESEARCH PROCESS 29
3.2.1 Choose the Texture Descriptor 29
3.2.2 Problem of Local Binary Pattern and Improvement 29
3.2.3 Problems of Traditional LBP When We Do Recognition 36
3.2.4 Make Sure Uniqueness 38
3.2.5 Get the Information of Spatial Distribution Based on LBP 39
3.2.6 Gaussian Mixture Models[9][25] [26] 40
3.2.7 Utilize Parameters of GMM to Get Features Against to Shift, Scale, and Rotation 43
3.3 CLASSIFIER 46
CHAPTER4 SYSTEM ARCHITECTURE AND EXPERIMENT RESULT 50
4.1 SYSTEM ARCHITECTURE 50
4.2 FACE DETECTION AND FACE IMAGE NORMALIZATION [28] 51
4.2.1 Integral Image 51
4.2.3 Cascaded Classifiers 54
4.2.3 Normalization of Facial Image and Pre-processing 56
CHAPTER5 CONCLUSION 67
5.1 CONCLUSION 67
REFERENCE 73

參考文獻
[1]M. A. Turk and A. P. Pentland, "Face recognition using eigenfaces," in Computer Vision and Pattern Recognition, 1991. Proceedings CVPR '91., IEEE Computer Society Conference on, 1991, pp. 586-591.
[2]L. Juwei, K. N. Plataniotis, and A. N. Venetsanopoulos, "Face recognition using LDA-based algorithms," Neural Networks, IEEE Transactions on, vol. 14, pp. 195-200, 2003.
[3]Z. Jie, J. Qiang, and G. Nagy, "A Comparative Study of Local Matching Approach for Face Recognition," Image Processing, IEEE Transactions on, vol. 16, pp. 2617-2628, 2007.
[4]M. Yoshida, T. Kamio and H. Asai, “Face Image Recognition by 2-Dimensional Discrete Walsh Transform and Multi-Layer Neural Network,” IEICE Trans. Fundamentals, Vol. E86-A, no.10, pp. 2623-2627, Oct. 2003.
[5]J. Xiao-Yuan and D. Zhang, "A face and palmprint recognition approach based on discriminant DCT feature extraction," Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on, vol. 34, pp. 2405-2415, 2004.
[6]H. K. Ekenel and R. Stiefelhagen, "Analysis of Local Appearance-Based Face Recognition: Effects of Feature Selection and Feature Normalization," in Computer Vision and Pattern Recognition Workshop, 2006. CVPRW '06. Conference on, 2006, pp. 34-34.
[7]G. Aguilar-Torres, G. Sanchez-Perez, M. Nakano-Miyatake, and H. Perez-Meana, "Face Recognition Algorithm Using the Discrete Gabor Transform," in Electronics, Communications and Computers, 2007. CONIELECOMP '07. 17th International Conference on, 2007, pp. 35-35.
[8]J. G. Daugman, "Complete discrete 2-D Gabor transforms by neural networks for image analysis and compression," Acoustics, Speech and Signal Processing, IEEE Transactions on, vol. 36, pp. 1169-1179, 1988.
[9]J. Ilonen, J. K. Kamarainen, P. Paalanen, M. Hamouz, J. Kittler, and H. Kalviainen, "Image Feature Localization by Multiple Hypothesis Testing of Gabor Features," Image Processing, IEEE Transactions on, vol. 17, pp. 311-325, 2008.
[10]R. M. Mutelo, W. L. Woo, and S. S. Dlay, "Discriminant analysis of the two-dimensional Gabor features for face recognition," Computer Vision, IET, vol. 2, pp. 37-49, 2008.
[11]L. Shu and A. C. S. Chung, "Texture Classification by using Advanced Local Binary Patterns and Spatial Distribution of Dominant Patterns," in Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on, 2007, pp. I-1221-I-1224.
[12]S.-W. Lee and S.Z. Li, “Learning Multi-scale Block Local Binary Patterns for Face Recognition”, ICB 2007, LNCS 4642, pp. 828–837, 2007.
[13]Xiao-Guang Lv, Jie Zhou, and Chang-Shui Zhang , “A novel algorithm for rotated human face detection”, IEEE Conference on Computer Vision and Pattern Recognition, Vol.1, pp.760-765, 2000.
[14]J. Choi, S. Kim, and P. Rhee, “Facial components segmentation for extraction facial feature”, Second International Conference on Audio- and Video-based Biometric Person Authentication, 1999.
[15]R. L. Hsu, M. Abdel-Mottaleb, and A. K. Jain, “Face Detection in Color Images”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.24, pp.696-706, 2002.
[16]H. A. Rowley, S. Baluja, T. Kanade, “Rotation Invariant Neural Network-Based Face Detection”, IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp.38-44, 1998.
[17]J. Terrillon, M. Sadek, H. Fukamachi, and S. Akamatsu, “Invariant face detection with support vector machines”, Proceedings of the 15th International Conference on Pattern Recognition, Vol.4, pp.210-217, 2000.
[18]P. Viola and M. Jones, "Rapid object detection using a boosted cascade of simple features," in Computer Vision and Pattern Recognition, 2001. CVPR 2001. Proceedings of the 2001 IEEE Computer Society Conference on, 2001, pp. I-511-I-518 vol.1.
[19]A.M. Martinez, A.C. Kak, “PCA versus LDA”, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 23, No. 2, pp. 228-233, 2001
[20]J. G. Daugman, “Complete discrete 2D Gabor transform by neural networks for image analysis and compression”, IEEE Trans. on Acoustic Speech and Signal Proc., vol. 36, No. 7, pp. 1169-1179, July 1988.
[21]W. S. Yambor, B. A. Draper, and J. R. Beveridge, “Analyzing PCA-based Face Recognition Algorithm: Eigenvector Selection and Distance Measures”,July 2000
[22]J.C. BURGES, "A Tutorial on Support Vector Machines for Pattern Recognition", Data Mining and Knowledge Discovery, 2, 121–167, 1998 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands
[23]V. N. Vapnik. Statistical learning theory. John Wiley & Sons, New York, 1998.
[24]T. Ojala, M. Pietikainen, and T. Maenpaa, "Multiresolution gray-scale and rotation invariant texture classification with local binary patterns," Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 24, pp. 971-987, 2002.
[25]“Gaussian Mixture Model.”[Online].Available: http://neural.cs.nthu.edu.tw/jang/books/dcpr/doc/08gmm.pdf
[26]D. Reynolds, T. Quatieri and R. Dunn, Speaker verification using adapted Gaussian mixture models. Digital Signal Process. 10 (2000), pp. 19–41
[27]G. Guodong, S. Z. Li, and C. Kapluk, "Face recognition by support vector machines," in Automatic Face and Gesture Recognition, 2000. Proceedings. Fourth IEEE International Conference on, 2000, pp. 196-201.
[28]Real-Time face recognition and verification system base on Rapid Object Detection,2007
[29]L.-F. Chen, H.-Y.M. Liao, J.-C. Lin, C.-C. Han, “Why Recognition in a Statistics-based Face Recognition System Should be based on the Pure Face Portion: a Probabilistic Decision-based Proof”, Pattern Recognition, Vol.34, No.5, 2001, pp. 1393-1403
[30]C. Wong, D.Kortenkam, and M. Speich, “A mobile robot that recognizes people”, IEEE Conference on Tools with Artificial Intelligence, 1995.
[31]http://www.sourceforge.net/projects/opencvlibrary
[32]C.-C. Chang and C.-J. Lin. LIBSVM: a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/˜cjlin/libsvm.
[33]“The orl database offaces.” [Online]. Available: http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html
[34]Q. Xipeng and W. Lide, "Face recognition by stepwise nonparametric margin maximum criterion," in Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on, 2005, pp. 1567-1572 Vol. 2.
[35]Z. Lichun, C. Junwei, L. Yue, and P. Wang, "Face Recognition Using Scale Invariant Feature Transform and Support Vector Machine," in Young Computer Scientists, 2008. ICYCS 2008. The 9th International Conference for, 2008, pp. 1766-1770.
[36]M. Sujaritha and S. Annadurai, "Face Recognition Using Wavelet Transform and Locally Discriminating Projection," in Conference on Computational Intelligence and Multimedia Applications, 2007. International Conference on, 2007, pp. 436-440.
[37]M. M. Khan, M. Y. Javed, and M. A. Anjum, "Face Recognition using Sub-Holistic PCA," in Information and Communication Technologies, 2005. ICICT 2005. First International Conference on, 2005, pp. 152-157.
[38]W. Jianguo, Y. Wankou, Y. Hui, and Y. Jingyu, "ICA Based Minimum Discriminant Analysis and Its Application to Face Recognition," in Pattern Recognition, 2008. CCPR '08. Chinese Conference on, 2008, pp. 1-4.
[39]Y. Xuelian, W. Xuegang, and L. Benyong, "A Direct Kernel Uncorrelated Discriminant Analysis Algorithm," Signal Processing Letters, IEEE, vol. 14, pp. 742-745, 2007.
[40]Z. Yu-jie, Y. Jing-yu, Y. Jian, W. Xiao-jun, and Y. Dong-jun, "A Complete and Rapid Feature Extraction Method for Face Recognition," in Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, 2006, pp. 469-472.