摘要
人脸定位是人脸识别的第一步。定位的好坏直接决定着后续识别的准确程度。由于人脸的角度,遮蔽和阴影等非线性因素导致了多角度人脸识别比固定角度(例如正面)的人脸识别困难得多。
本文就两种人脸识别方法分别利用不同的数学工具设计了两种不同的分类器。
针对正面人脸识别的分类器使用两步定位法,先进行粗定位再进行细定位。粗定位算法的主要步骤是,先对图像进行预处理,再使用OAC滤波器滤波,最后使用阈值分割法将图像分割为人脸区域、可能人脸区域和非人脸区域三部分。为人脸细定位算法减少了计算量,增加了准确率。对CMU数据库中的图像进行测试,非人脸区域的排除率达到99%,误检率仅有1.3%。实验结果证明了该算法的有效性和可靠性。
针对多角度人脸识别,角度空间被分为多个部分。每部分设计一种分类器。无论是否是人脸,都要使用支持向量回归进行人脸角度估计。人脸角度估计的结果用于分类器来估计是否是人脸。由于姿势估计这一步,大大减少了计算量。同时,由于每个分类器都只针对一部分角度的人脸,所以提高了准确率。经过对比,我所用的算法同现存人脸定位方法中速度较快的特征脸法相比,具有同样的速度,而且具有更高的准确率。
Detection is the first step of face recognition, which directly affects the result of face recognition. Detecting faces across multiple views is more challenging than in a fixed view, e.g. frontal view, owing to the significant non-linear variation caused by rotation in depth, self-occlusion and self-shadowing.
In this paper, I apply two classifier using two different methods, one for frontal view detection and the other for multiple views.
The first method uses two-step classifier. The first step is rough detection, and the second step is delicate detection. The main processes of the rough detection algorithm are that, firstly, preprocess the image, secondly, filter in Fourier transform domain by OAC, thirdly, segment the image into "face" area, "may-be face" area and "none-face area". The testing on CMU face database shows that the algorithm successfully segments the image with "non-face" rejecting rate of 99% and false alarm rate of 1.3%. The experiment results confirm that the algorithms are efficient and reliability.
For the second method the view sphere is separated into several small segments. On each segment, a face detector is constructed. We explicitly estimate the pose of an image regardless of whether or not it being a face. A pose estimator is constructed using Support Vector Regression. The pose information is used to choose the appropriate face detector to determine if it is a face. With this pose-estimation based method, considerable computational efficiency is achieved. Meanwhile, the detection accuracy is also improved since each detector is constructed on a small range of views. We developed a novel algorithm for face detection by combining the Eigenface and SVM methods which performs almost as fast as the Eigenface method but with a significant improved speed.
引文
[1]李刚,高政,人脸识别方法综述,计算机应用,2003年08期,4-9页
[2]Shakhnarovich,G.,Viola,P.,and Moghaddam,B..A unified learning framework for real time facedetection and classification. In Proc.IEEE Int'l.Conf.Automatic Face and Gesture Recognition,2002, pp.14-21
[3]Baek K, Draper B A, Beveridge J R, et al. PCA vs. ICA:A Comparison on the FERET Data Set. CVPRIP in Conjunction with the 6thJCIS, Durham, North Carolina, 2002-03-08
[4]Samaria F, Harter A. Parametrisation of a Stochastic Model for Human Face Identification. In:Proceedings of IEEE Workshop on Applica- tions on Computer Vision,1994
[5]Lu X, Wang Y, Jain A K. Combining Classifiers for Face Recognition. Proc. ICME 2003, IEEE International Conference on Multimedia & Expo, Baltimore, MD, 2003-07-06, Ⅲ:13-16
[6]M.-H.Yang, D.Kriegman, and N. Ahuja, "Detecting faces in images:a survey", TPAMI,24(1):34-58,2002.
[7]G. Yang and T. S. Huang, "Human Face Detection in Complex Background," Pattern Recognition, vol.27, no.1, pp.53-63,1994.
[8]S. McKenna, S. Gong, and Y. Raja, "Modelling Facial Colour and Identity with Gaussian Mixtures," Pattern Recognition, vol.31, no.12, pp.1883-1892,1998.
[9]K.-K. Sung and T. Poggio, "Example-Based Learning for View-Based Human Face Detection," IEEE Trans. Pattern Analysis and Machine Intelligence, vol.20, no.1, pp. 39-51, Jan.1998.
[10]A. Rajagopalan, K. Kumar, J. Karlekar, R. Manivasakan, M. Patil,U. Desai, P. Poonacha, and S. Chaudhuri, "Finding Faces inPhotographs," Proc. Sixth IEEE Int'l Conf. Computer Vision, pp.640-645,1998.
[11]J. Miao, B. Yin, K. Wang, L. Shen, and X. Chen, "A Hierarchical Multiscale and Multiangle System for Human Face Detection in a Complex Background Using Gravity-Center Template," Pattern Recognition, vol.32, no.7, pp.1237-1248,1999.
[12]M.Elad, Y. Hel-Or and R. Keshet," Rejection based classifier for face detection", Pattern Recognition Letters,23,2002,1459-1471.
[13]L. Yaroslavsky and M. Eden, Fundamentals of Digital Optics, Birkhauser, Boston, 1996.
[14]B.-Z. Shdick and L.P. Yaroslavsky, "Object localization using linear adaptive filters", Inter.Workshop on Vision, Modeling and Visualization,Stuttgart,2001, pp. 11-18.
[15]刘艳丽,赵跃龙,人脸识别技术研究进展,计算机工程,2005年2月第3期,10-13页
[16]赵丽红,刘继红,徐新和,人脸检测算法综述,计算机应用研究,2004年第9期,1-4页
[17]T.S.Huang,G.Z.Yang.Human Face Detection in a complex Background, Pattern Recognition,1994,27(l):53-63
[18]Kotropoulos,Pitas,Rule-based Face Detection in Frontal Views Proc.Int'lConf.Acoustics,Speech and Signal Processing,1997,2537-2540
[19]李华胜,杨桦,袁保宗.人脸识别系统中的特征提取,.北京交通大学学报,2001,25(2)：4.
[20]姜军,张桂林.一种基于知识的快速人脸检测方法[J].中国图像图形学报(A版),2002,7(1).
[21]卢春雨,张长水,闻方,等.基于区域特征的快速人脸检测法[J].清华大学学报(自然科学版),1999,39(1)：101-105.
[22]S.A.Sirohey.Human Face Segmentation and Identification Technical Report.CS-TR-3176[R] University of Maryland,1993
[23]H.P.Graf.Locating Faces and Facial Parts[C].Proc.1st Int'Workshop Automatic Face and Gesture Recognition.1995,41-46
[24]T.K.Leung,M.C.Burl,P.Prona.Finding Faces in Clutters Scenes Using Random Label Graph Matching.Proc.5thIEEE Int'lConf.Comupter Vision.1995,637-644
[25]K.C.Yow,R.Cipolla.A Probabilistic Framework for Perceptual Grouping of Features for Human Face Detection.Prc.2nd Int'l ConfAutomatic Face and Gesture Recognition,1996,16-20
[26]K.C.Yow.R Cipolla.Enbancing Human Face Detection Using Motion and Active Contous.3rd.Asian Conf.Computer Vision.1998,515-522
[27]Platt J. Fast Training of Support Vector Machines Using Sequential Minimal Optimization[A]. Scholkopf B, Burges C, Smola A, Advances in Kernel Methods2 Support Vector Learning[C]. Cambridge MA:MIT Press,1999.
[28]Fei S, Lawrence K, Daniel D. Multiplicative updates for non negative quadratic programming in support vector machines[A].Suzanna B, Sebastian T, Klaus O. Advances in Neural Information Processing Systems[C].Cambridge M A:MIT Press,2003.
[29]Gert C, Tomaso P. Incremental and decremental support vector machine learning[EB/OL]. http://citeseer.ist.psu.edu/cauwen-berghs00incremental. html, 2000-01-01/2005-05-13.
[30]Edgar O. Training Support Vector Machines:An Application to Face Detection [EB/OL].http://citeseer.ist.psu.edu/osuna97training.html,1997-01-01/2005-05-13.
[31]Chang C C, Hsu C W, Lin C J. The analysis of decomposition methods for support vector machines[EB/OL]. http://citeseer.ist.psu.edu/327075.html, 1999-01-01/2005-05-13.
[32]Marcelo B A, Antonio P B, Joao P B. SVM-KM:Speeding SVMs Learning with a priori Cluster Selection and k-Means[EB/OL]. http://csdl. computer. org/comp/ proceedings/sbrn/2000/0856/00/08560162abs. htm,2000-01-22/2005-05-13.
[33]Jinbo B, Bennett K, Embrechts M. Dimensionality Reduction via Sparse Support Vector Machines [J].Journal of Machine Learning Research,2003,3(1):1229-1243.
[34]Lin C J. Asymptotic convergence of an SMO algorithm without any assumptions [J]. IEEE Transactions on Neural Networks,2001,13(1):248-250.
[35]Friess T T. Thekernel-adatron algorithm:A fast and simple learning procedure for support vector machines[EB/OL]. http://citese2er.ist.psu.edu/284272.html,1998-01-01/2005-05-13.
[36]Vijayakumar S, Wu S. Sequential Support Vector Classifiers and Regression[EB/OL].http://citeseer.ist. psu.edu/284272.html,1998-01-01/2005-05-13.
[37]Mangasarian O L, Musicant D R. Successive over relaxation for support vector machines [J]. IEEE Transactionson Neural Networks
[38]Mario M.On-line Support Vector Machines for Function Approximation [EB/OL]. http://www.lsi.upc.es/-mmartin/svmr.html,2002-03-10/2005-05-13.
[39]Vapnik V, Golowich S, Smola A. Support vector method for function approximation,regression estimation, and signal processing[A].Mozer M, Jordan M, Petsche T. Advances in Neural Information Processing Systems [C].CambridgeMA:MITPress,1997.
[40]李建民,张钹林,福宗.支持向量机的训练算法[J].清华大学学报(自然版),2003,43(1)：120-124.
[41]张浩然,汪小东,支持向量机的学习方法综述,浙江大学学报(自然科学版), 2005年8月,第3期
[42]Cristianini N, Shawe-Tayor J. An introduction to support vector Machines and other kernel-based learning methods.Cam2 bridge:Cambridge University Press,2000.
[43]齐志泉等,支持向量机中的核函数选择问题,控制工程,2005年7月第4期,379-381页
[44]Yongming Li,Shaogang Gong,Support Vector Machine based Multi-view Face Detection and Recognition, Image vision computing,2004(22)413-427
[45]ICraw, HEllis, JLishman. Automatic Extraction of Face Features. Pattern Recognition Letters,1987,5:183-1871
[46]VGovindaraju. Locating Human Faces in Photographs. Int.l Computer Vision,1996, 19(2):129-1461
[47]JMiao,etal.A Hierarchical Multi scale and Multi angle System for Human Face Detection in a Complex Background Using Gravity-Center Template. Pattern Recognition,1999,32(7):1237-12481
[48]梁路宏,艾海舟,何克忠.基于多模板匹配的单人脸检测[J].中国图像图形学报,1999,823-8301
[49]周激流,张晔,等.基于先验模板的人脸面部特征提取的研究.计算机辅助设计与图形学学报,2000,12(5)：337-3391
[50]M Kirby, L Sirovich. Application of the Karhunen-Loe-ve Procedure for The Characterization of Human Faces[J].IEEE Trans. Pattern Analysis and Machine Intelligence,1990,12(1):103-1081
[51]MTurk, A Pentland. Eigenfaces for Recognition. Cognitive Neuro-science, 1991,3(1):71-86.
[52]K-K Sung, T Poggio. Example-based Learning for View-based Human Face Detection.Technical Report AI Memo1521, Massachusetts Inst. Of Technology AILab,1994.
[53]M-H Yang, N Ahuja, D Kriegman. Mixtures of Linear Subspaces for Face Detection.4thInt.lConf.AutomaticFaceandGestureRecognition,2000.70-761
[54]T Agui, Y Kokubo, etal. Extraction of Face Recognition from Monochro-Matic Photographs Using Neural Networks.2ndInt.lConf.Automation, Robotics, and Computer Vision,1992.
