基于独立成分分析的人脸识别方法研究
|
摘要
人脸识别技术是利用计算机分析人脸图像,从中获取有效识别信息,用来辨认身份的一门技术。与其它生物识别技术相比,人脸识别具有直接、友好、方便、非侵犯性、稳定等优点,因而有着广阔的应用前景。本文针对一些已经存在的人脸识别方法进行改进,推动了人脸识别技术的进一步发展。
针对人脸图像维数较大,直接进行处理时计算复杂度较大的问题,本文利用小波分解对人脸图像进行了预处理。取经过小波分解后的低频部分作为预处理后的图像,降低了图像的维数,去除了部分高频噪声,在不影响识别精度的同时降低了训练时的计算复杂度。
针对主成分分析方法(PCA)只能消除图像二阶冗余的缺点,本文利用盲信号处理中的高阶数据处理方法——独立成分分析方法进行人脸图像的特征提取,不仅提取了图像的二阶信息,也充分利用了人脸图像的高阶统计特性,提取的特征能更好地揭示人脸图像的本质特征,为进一步分类打好基础。
分类识别阶段,由于人脸特征具有小样本、非线性及高维模式等特点,而支持向量机(SVM)在解决这类问题时表现出许多特有的优点,所以,本文将支作持向量机为分类器应用于人脸识别,与简单的距离分类器相比,大大提高了分类效果。
最后,本文基于ORL人脸库对各种算法进行了计算机仿真实验,给出了具体的实验数据及结果分析。
The face recognition technology is to analyze face images with computer, obtain the valid information and realize the identification of human faces. Compares with other biometrics, the face recognition technology is direct, friendly, convenient, non-infringe and stable, therefore, the face recognition technology has an extremely broad application prospect. In this paper, some existing methods for face recognition are improved and face recognition technology is promoted.
For the problem that the dimension of face images is large and it is difficult to deal with such large images directly with computer, in this paper, wavelet decomposition is used as a pre-processing method. The result of the wavelet decomposition is that the dimension of the images is reduced and some high-frequency noise is filtered. Thus, the training computational complexity is reduced without affecting the identification accuracy.
Because of the shortcomings of the Principal Component Analysis (PCA) that it can only eliminate the second-order redundancy in the images, in this paper, Independent Component Analysis (ICA)-a kind of high-order data processing method in Blind Signal Processing (BSP) field, is used to extract the features of the face images. The ICA method not only extracts the second-order information between images, but also makes full use of the higher-order statistical features of the face images. The features extracted by ICA method can reveal the nature of the face images better. As a result, it makes a good foundation for further classification.
In classification stage, the facial feature is characteristic of small sample, nonlinearity and high-dimensional model, and Support Vector Machine (SVM) shows a number of unique advantages in solving this kind of problems. So, SVM is applied to face recognition as a classifier in this paper. Compared with the simple distance classifiers, the classification results gotten by using SVM are significantly improved.
At last, based on ORL face database, the simulation experiment is realized by computer with a variety of algorithms. The experimental data and result analysis are given in detail.
针对人脸图像维数较大,直接进行处理时计算复杂度较大的问题,本文利用小波分解对人脸图像进行了预处理。取经过小波分解后的低频部分作为预处理后的图像,降低了图像的维数,去除了部分高频噪声,在不影响识别精度的同时降低了训练时的计算复杂度。
针对主成分分析方法(PCA)只能消除图像二阶冗余的缺点,本文利用盲信号处理中的高阶数据处理方法——独立成分分析方法进行人脸图像的特征提取,不仅提取了图像的二阶信息,也充分利用了人脸图像的高阶统计特性,提取的特征能更好地揭示人脸图像的本质特征,为进一步分类打好基础。
分类识别阶段,由于人脸特征具有小样本、非线性及高维模式等特点,而支持向量机(SVM)在解决这类问题时表现出许多特有的优点,所以,本文将支作持向量机为分类器应用于人脸识别,与简单的距离分类器相比,大大提高了分类效果。
最后,本文基于ORL人脸库对各种算法进行了计算机仿真实验,给出了具体的实验数据及结果分析。
The face recognition technology is to analyze face images with computer, obtain the valid information and realize the identification of human faces. Compares with other biometrics, the face recognition technology is direct, friendly, convenient, non-infringe and stable, therefore, the face recognition technology has an extremely broad application prospect. In this paper, some existing methods for face recognition are improved and face recognition technology is promoted.
For the problem that the dimension of face images is large and it is difficult to deal with such large images directly with computer, in this paper, wavelet decomposition is used as a pre-processing method. The result of the wavelet decomposition is that the dimension of the images is reduced and some high-frequency noise is filtered. Thus, the training computational complexity is reduced without affecting the identification accuracy.
Because of the shortcomings of the Principal Component Analysis (PCA) that it can only eliminate the second-order redundancy in the images, in this paper, Independent Component Analysis (ICA)-a kind of high-order data processing method in Blind Signal Processing (BSP) field, is used to extract the features of the face images. The ICA method not only extracts the second-order information between images, but also makes full use of the higher-order statistical features of the face images. The features extracted by ICA method can reveal the nature of the face images better. As a result, it makes a good foundation for further classification.
In classification stage, the facial feature is characteristic of small sample, nonlinearity and high-dimensional model, and Support Vector Machine (SVM) shows a number of unique advantages in solving this kind of problems. So, SVM is applied to face recognition as a classifier in this paper. Compared with the simple distance classifiers, the classification results gotten by using SVM are significantly improved.
At last, based on ORL face database, the simulation experiment is realized by computer with a variety of algorithms. The experimental data and result analysis are given in detail.
引文
[1]Jonathan Grudin. Three Faces of Human Computer Interaction [J], IEEE Annals of the History of Computing, Volume 27, Issue 4:46-62 Oct.. Dec.2005
[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]高全学,潘泉,程咏梅,张洪才,王勇.基于肤色和独立成分的人脸检测计算机应用研究,2004:160 ~162
[4]周激流,张晔,人脸识别理论研究进展,计算机辅助设计与图形学学报,1999,11(2):180~184
[5]Sir Francis Galton. Personal Identification and Subscription-1[J].Nature,1906, 1888(21):173~177
[6]苏剑波,徐波著.应用模式识别导论[M],上海:上海交通大学出版社,2001,5~8
[7] Chellappa R,Wilson C L,Sirohey S. Human and machine recognition of faces:a survey.Proc.IEEE.1995,83(5):704-740
[8]Goldstein, Harmon, Lesk. Identification of human faces, Proceedings of the IEEE, 1971, 59(5):748~760
[9]Y. Kaya and K. Kobayashi, "A Basic Study on Human Face Recognition, "Frontiers of Pattern Recognition, S. Watanabe, ed., 1972, p265
[10]Kanade.T. Geometrical aspects of interpreting images as a three-dimensional scene, Proceedings of the IEEE, 1983, 71(7):789~802
[11]Brunelli R, Poggio T, Face Recognition: Features vs. Templates, IEEE Transactions on Pattern Anlysis and Machine Intelligence, 1993, 15(10):1042~1052
[12]Kotropoulos C, Tefas A, Pitas I. Frontal face authentication using morphological elastic graph matching, IEEE Transactions on Image. Processing, 2000, 9(4):555~560
[13]Turk M, Pentland A P, Face recognition using eigfaces in Proc. IEEE Conf. On Computor vision and pattern Recognition, 1991, pp.586~591
[14]Bulhumeur P N, Hespanha J P, Kriegman D J.Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Liear Projection, IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997, 19(7):711~720
[15]M.S.Bartlett, H.M.Lades, T.J.Sejnowski, Independent component representation for face recognition .Proc. of SPIE, 2399, 1998:528~539
[16]O.Deniz, M.Castrillon, M.Hernandez, Face recognition using independent component analysis and support vector machines. Int. Conf, On Audio- and Video-Based Person Authentication. Lecture Notes, 2001: 59~64
[17]Valentin D, Abdi H, O’Toole A J, et al. Connectionist Models of Face Processing: A Survey. Pattern Recognition, 1994, 27(9): 1209~1230
[18]Samaria F, Young S.A HMM-Based Architecture For Face Identification. Image and Vision Computing, 1994, 12(8):537~543
[19]P.J.Phillips, H.Moon, etc.“The FERET Evaluation Methodology for Face-Recognition Algorithms”, IEEE Transactions on PAMI, Vol.22, No.10, pp1090-1104, 2000
[20]Birchfield S, An elliptical head tracker. In 31st Asilomar Conference on Signals, System and Computers, 1997, pp.1710~1714
[21]M. Lades, J.C. Vorbruggen, J. Buhmann, etc, Distortion invariant object recognition in the dynamic link architecture,IEEE Trans. On Computers, 1993, 42(3):300~311
[22]M.Turk, A. Pentland, Eigfaces for recognition. J. Cognitive Neuroscience, 1991, 3(1):71~86
[23] Chellappa R.,Wilson C.L.,Sirohey S.Humaa and Machine Recognition of Faces:a Survey[C],Proceedings of the IEEE,1995,83(5):705~741
[24]K.Pearson. On lines and planes of closest fit to systems of points in space. Philosophical Magazine, 1901, 2:559~572
[25]M. Kirby, L. Sirovich, Application of the Karhunen-Loeve procedure for the characterization of human faces, IEEE Transactions on Pattern Analysis and Machine Intelligence 12 (1990) 103~107
[26]L. Sirovich, M. Kirby, A low-dimensional procedure for the characterization of human faces, Journal of the Optical Society of America 4 (1987) 519~524
[27]G.W. Cottrell and M.K. Fleming, Face recognition using unsupervised feature extraction, presented at International Neural Network Conference, Dordrecht, 1990
[28]P Common, Independent component analysis, a new concept? [J], Signal Processing, 1994, 36(3): 287~314
[29]A.Hyvarinen, Survey on independent component analysis [J], Neural Computing Surveys, 1999, 2(4):94~128
[30]夏良正著,数字图像处理[M],南京:东南大学出版社, 1999, 135~142
[31]杨福生著,小波变换的工程分析与应用[M],北京:科学出版社,2000,110~195
[32]A.Hyvarinen and E.Oja, Independent component analysis: a tutorial [J], Neural Networks 2000, 13(4):411~430
[33]L.Tong, Y.Inouye, and R.Liu, Waveform preserving blind estimation of multiple independent sources, IEEE Trans. on Signal Processing, 41: 2461~2470, 1993
[34]J.Hreault, et al. Space or time adaptive signal processing by neural network models, in Denkered.“Neural Network for Computing”. AIP Conference Processing, American Institute of Physics, 151, 1986
[35]C. Jutten, et al. Blind separation of sources, Pt.?: An adaptive algorithm based on neuromimetic architecture. Signal Processing, 24(1), 1~10, 1991
[36]P.Commen, et al. Blind separation of sources, Pt.: Problem statement. Signal Processing, 24(1), 11~20, 1991
[37]A.J.Bell, T.J.Sejnowski, An information maximization approach to blind separation and blind disconsolation.Neural Computation.1995, 7(2): 1129~1159
[38]T.W.Lee, M. Girolami, T.Sejnowski, Independent component analysis using an extended infomax algorithm for mixed sub-gaussian and super-gaussian source, Neural, Computation, 1999, 11: 417~441
[39]Amari, S.L.Cichoicki.A. & Yang.H, A new learning algorithm for blind signal separation, Advances in Neural Information Processing Systems, 1996, 8(3):757~763
[40]A.Hyvarinen. Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans. on Neural Networks, 10(3): 626~634, 1999
[41]Gaeta.M, Lacoume.J.L. 1990, Sources separation without a priori know1edge: the maximum Likelihood solution, in Proc. EUSIPCO, Barcelona, Spain, pp621~624
[42]J.F.Cardoso. Higher order contrast for independent component analysis. Neural Computing, 11(1):157~193, 1999
[43]J.F.Cardoso, Blind beamforming for non-Gaussian signal, IEEE-Pro.F, 140(6): 362~370, 1993
[44]G.Deco, et al.. Information maximization and independent component analysis, Is there a difference? Neural Computation, 10(8), 2085~2101,1998
[45]Aapo. Hyvarinen, Erkki.oja, Independent component analysis: algorithms and applications. Neural Networks, 2000, 13(3):258~263
[46]田陈魁,应用概率统计,北京:清华大学出版社,2000
[47]Bell A.J, Sejnowski T.J, The“independent components”of natural scenes are edge filters, Vision Res. Vol. 37, no. 23, pp. 3327~3338, 1997
[48]David J. Field,“What is the goal of sensory coding?”, Neural Computation, vol.6, pp. 559~601, 1994
[49] A. V. Oppenheim and J. S. Lim,“The importance of phase in signals,”Proc. IEEE, vol. 69, pp. 529~541, 1981
[50] L. N. Piotrowski and F. W. Campbell,“A demonstration of the visual importance and flexibility of spatial-frequency, amplitude and phase,”Perception, vol.11, pp. 337~346, 1982
[51]Moghaddam B.Jebara T.Pentland, A Bayesian face recognition Pattern recognition. 2000, 33 1171~1782
[52] C.J.C. Burges, A tutorial on support vector machines for pattern recognition Data Mining Knowledge Discovery, 1998, 2(2) 121~167
[53]Czeminski R.Yasn A.Hartsough D Use of support vector machines in pattern classification application to QSAR studies Quart Struct-Act Relat. 2001(20)227~240
[54]Cristianim N, Shawe-Taylor J, An Introduction to Support Vector Machines and Other Kernal-Based Learning Methods Publishing House Of Electronics Industry 2004
[55]M. S. Bartlett, J. R. Movellan, T. J. Sejnowski. Face Recognition by Independent Component Analysis, IEEE Trans. on Neural Networks, 2002, 13(6):1450~1463
[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]高全学,潘泉,程咏梅,张洪才,王勇.基于肤色和独立成分的人脸检测计算机应用研究,2004:160 ~162
[4]周激流,张晔,人脸识别理论研究进展,计算机辅助设计与图形学学报,1999,11(2):180~184
[5]Sir Francis Galton. Personal Identification and Subscription-1[J].Nature,1906, 1888(21):173~177
[6]苏剑波,徐波著.应用模式识别导论[M],上海:上海交通大学出版社,2001,5~8
[7] Chellappa R,Wilson C L,Sirohey S. Human and machine recognition of faces:a survey.Proc.IEEE.1995,83(5):704-740
[8]Goldstein, Harmon, Lesk. Identification of human faces, Proceedings of the IEEE, 1971, 59(5):748~760
[9]Y. Kaya and K. Kobayashi, "A Basic Study on Human Face Recognition, "Frontiers of Pattern Recognition, S. Watanabe, ed., 1972, p265
[10]Kanade.T. Geometrical aspects of interpreting images as a three-dimensional scene, Proceedings of the IEEE, 1983, 71(7):789~802
[11]Brunelli R, Poggio T, Face Recognition: Features vs. Templates, IEEE Transactions on Pattern Anlysis and Machine Intelligence, 1993, 15(10):1042~1052
[12]Kotropoulos C, Tefas A, Pitas I. Frontal face authentication using morphological elastic graph matching, IEEE Transactions on Image. Processing, 2000, 9(4):555~560
[13]Turk M, Pentland A P, Face recognition using eigfaces in Proc. IEEE Conf. On Computor vision and pattern Recognition, 1991, pp.586~591
[14]Bulhumeur P N, Hespanha J P, Kriegman D J.Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Liear Projection, IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997, 19(7):711~720
[15]M.S.Bartlett, H.M.Lades, T.J.Sejnowski, Independent component representation for face recognition .Proc. of SPIE, 2399, 1998:528~539
[16]O.Deniz, M.Castrillon, M.Hernandez, Face recognition using independent component analysis and support vector machines. Int. Conf, On Audio- and Video-Based Person Authentication. Lecture Notes, 2001: 59~64
[17]Valentin D, Abdi H, O’Toole A J, et al. Connectionist Models of Face Processing: A Survey. Pattern Recognition, 1994, 27(9): 1209~1230
[18]Samaria F, Young S.A HMM-Based Architecture For Face Identification. Image and Vision Computing, 1994, 12(8):537~543
[19]P.J.Phillips, H.Moon, etc.“The FERET Evaluation Methodology for Face-Recognition Algorithms”, IEEE Transactions on PAMI, Vol.22, No.10, pp1090-1104, 2000
[20]Birchfield S, An elliptical head tracker. In 31st Asilomar Conference on Signals, System and Computers, 1997, pp.1710~1714
[21]M. Lades, J.C. Vorbruggen, J. Buhmann, etc, Distortion invariant object recognition in the dynamic link architecture,IEEE Trans. On Computers, 1993, 42(3):300~311
[22]M.Turk, A. Pentland, Eigfaces for recognition. J. Cognitive Neuroscience, 1991, 3(1):71~86
[23] Chellappa R.,Wilson C.L.,Sirohey S.Humaa and Machine Recognition of Faces:a Survey[C],Proceedings of the IEEE,1995,83(5):705~741
[24]K.Pearson. On lines and planes of closest fit to systems of points in space. Philosophical Magazine, 1901, 2:559~572
[25]M. Kirby, L. Sirovich, Application of the Karhunen-Loeve procedure for the characterization of human faces, IEEE Transactions on Pattern Analysis and Machine Intelligence 12 (1990) 103~107
[26]L. Sirovich, M. Kirby, A low-dimensional procedure for the characterization of human faces, Journal of the Optical Society of America 4 (1987) 519~524
[27]G.W. Cottrell and M.K. Fleming, Face recognition using unsupervised feature extraction, presented at International Neural Network Conference, Dordrecht, 1990
[28]P Common, Independent component analysis, a new concept? [J], Signal Processing, 1994, 36(3): 287~314
[29]A.Hyvarinen, Survey on independent component analysis [J], Neural Computing Surveys, 1999, 2(4):94~128
[30]夏良正著,数字图像处理[M],南京:东南大学出版社, 1999, 135~142
[31]杨福生著,小波变换的工程分析与应用[M],北京:科学出版社,2000,110~195
[32]A.Hyvarinen and E.Oja, Independent component analysis: a tutorial [J], Neural Networks 2000, 13(4):411~430
[33]L.Tong, Y.Inouye, and R.Liu, Waveform preserving blind estimation of multiple independent sources, IEEE Trans. on Signal Processing, 41: 2461~2470, 1993
[34]J.Hreault, et al. Space or time adaptive signal processing by neural network models, in Denkered.“Neural Network for Computing”. AIP Conference Processing, American Institute of Physics, 151, 1986
[35]C. Jutten, et al. Blind separation of sources, Pt.?: An adaptive algorithm based on neuromimetic architecture. Signal Processing, 24(1), 1~10, 1991
[36]P.Commen, et al. Blind separation of sources, Pt.: Problem statement. Signal Processing, 24(1), 11~20, 1991
[37]A.J.Bell, T.J.Sejnowski, An information maximization approach to blind separation and blind disconsolation.Neural Computation.1995, 7(2): 1129~1159
[38]T.W.Lee, M. Girolami, T.Sejnowski, Independent component analysis using an extended infomax algorithm for mixed sub-gaussian and super-gaussian source, Neural, Computation, 1999, 11: 417~441
[39]Amari, S.L.Cichoicki.A. & Yang.H, A new learning algorithm for blind signal separation, Advances in Neural Information Processing Systems, 1996, 8(3):757~763
[40]A.Hyvarinen. Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans. on Neural Networks, 10(3): 626~634, 1999
[41]Gaeta.M, Lacoume.J.L. 1990, Sources separation without a priori know1edge: the maximum Likelihood solution, in Proc. EUSIPCO, Barcelona, Spain, pp621~624
[42]J.F.Cardoso. Higher order contrast for independent component analysis. Neural Computing, 11(1):157~193, 1999
[43]J.F.Cardoso, Blind beamforming for non-Gaussian signal, IEEE-Pro.F, 140(6): 362~370, 1993
[44]G.Deco, et al.. Information maximization and independent component analysis, Is there a difference? Neural Computation, 10(8), 2085~2101,1998
[45]Aapo. Hyvarinen, Erkki.oja, Independent component analysis: algorithms and applications. Neural Networks, 2000, 13(3):258~263
[46]田陈魁,应用概率统计,北京:清华大学出版社,2000
[47]Bell A.J, Sejnowski T.J, The“independent components”of natural scenes are edge filters, Vision Res. Vol. 37, no. 23, pp. 3327~3338, 1997
[48]David J. Field,“What is the goal of sensory coding?”, Neural Computation, vol.6, pp. 559~601, 1994
[49] A. V. Oppenheim and J. S. Lim,“The importance of phase in signals,”Proc. IEEE, vol. 69, pp. 529~541, 1981
[50] L. N. Piotrowski and F. W. Campbell,“A demonstration of the visual importance and flexibility of spatial-frequency, amplitude and phase,”Perception, vol.11, pp. 337~346, 1982
[51]Moghaddam B.Jebara T.Pentland, A Bayesian face recognition Pattern recognition. 2000, 33 1171~1782
[52] C.J.C. Burges, A tutorial on support vector machines for pattern recognition Data Mining Knowledge Discovery, 1998, 2(2) 121~167
[53]Czeminski R.Yasn A.Hartsough D Use of support vector machines in pattern classification application to QSAR studies Quart Struct-Act Relat. 2001(20)227~240
[54]Cristianim N, Shawe-Taylor J, An Introduction to Support Vector Machines and Other Kernal-Based Learning Methods Publishing House Of Electronics Industry 2004
[55]M. S. Bartlett, J. R. Movellan, T. J. Sejnowski. Face Recognition by Independent Component Analysis, IEEE Trans. on Neural Networks, 2002, 13(6):1450~1463