人脸图像主流代数鉴别特征提取与融合方法研究
|
摘要
人脸识别是计算机视觉领域里的一个非常热门的研究课题,具有极高的科研价值和广阔的应用前景。而特征提取又是人脸识别中最关键的环节之一,直接影响到识别的准确率。
本文首先介绍了主成分分析、Fisher线性鉴别分析和最大散度差鉴别分析三种基于代数特征的人脸图像特征提取方法,并分析了三种方法的特点和弊端。
针对传统的线性分析方法中都需要求解平均样本的共性,本文提出了两种改进方法。经过在标准人脸图像库上的实验,证实了这两种方法都较好地去除了干扰样本对平均样本的影响,能有效地提高识别的准确率。
本文还尝试利用典型相关分析方法对改进后的Fisher线性鉴别分析方法和最大散度差鉴别分析方法进行特征融合,并用实验验证了所尝试的方法具有较高的识别率,达到了融合的效果。
Face recognition is a top research subject in the computer visual field. It has a very high scientific value and broad application prospects, and feature extraction is one of the most critical aspects in the area of face recognition, it affects the accuracy of identification directly.
This article describes three feature extraction methods for face image based on algebraic feature. They are principal component analysis、Fisher linear discriminant analysis、maximum scatter difference discriminant analysis. This article also analysis the features and disadvantages of the three methods.
According to the common needs of the average sample solution in traditional linear analysis methods, this article proposes two improved methods. After a face database in the standard experiments, it proves that both the two way can remove the influence of Average samples to Interference samples, also can improve the recognition accuracy.
This article also tries to use the canonical correlation analysis to fuse the features of the improved Fisher linear discriminant analysis and maximum scatter difference discriminant analysis, and through the experiments, it proved that the way we have tried has the high recognition rate, achieved the fusion results.
本文首先介绍了主成分分析、Fisher线性鉴别分析和最大散度差鉴别分析三种基于代数特征的人脸图像特征提取方法,并分析了三种方法的特点和弊端。
针对传统的线性分析方法中都需要求解平均样本的共性,本文提出了两种改进方法。经过在标准人脸图像库上的实验,证实了这两种方法都较好地去除了干扰样本对平均样本的影响,能有效地提高识别的准确率。
本文还尝试利用典型相关分析方法对改进后的Fisher线性鉴别分析方法和最大散度差鉴别分析方法进行特征融合,并用实验验证了所尝试的方法具有较高的识别率,达到了融合的效果。
Face recognition is a top research subject in the computer visual field. It has a very high scientific value and broad application prospects, and feature extraction is one of the most critical aspects in the area of face recognition, it affects the accuracy of identification directly.
This article describes three feature extraction methods for face image based on algebraic feature. They are principal component analysis、Fisher linear discriminant analysis、maximum scatter difference discriminant analysis. This article also analysis the features and disadvantages of the three methods.
According to the common needs of the average sample solution in traditional linear analysis methods, this article proposes two improved methods. After a face database in the standard experiments, it proves that both the two way can remove the influence of Average samples to Interference samples, also can improve the recognition accuracy.
This article also tries to use the canonical correlation analysis to fuse the features of the improved Fisher linear discriminant analysis and maximum scatter difference discriminant analysis, and through the experiments, it proved that the way we have tried has the high recognition rate, achieved the fusion results.
引文
[1] Robert W.Frischholz,Ulrich Diecklmann.BioID:A Multimodal Biometric Identification System.IEEE Computer.2000,33.2:64-68.
[2] Bertillon and J.Ruiz Solar. Pattern Recognition with Support Vector Machines volume 2388,1971,chapter On the Generalization of Kernel Machines,pages 24-39.Springer,August.
[3] Allen A L.Personal Descriptions,Butterworth,1950.
[4] Parker F I.Computer Generated Animation of Faces. In Processing ACM ann Conference, 1972, l:451-457.
[5] Goldation R J,Harmon L D,Lesk A B.Man manchine interaction in human face Identification. Bell Syst Tech Journal,1973,52:265-289.
[6] Kaya Y Kobayashi K.A Basic study on Human Recognition. In Frontiers of attern Recognition. New York:Academic,1971:265-289.
[7] Kanad T. Picture processing system by computer and recognition of human face [Ph D Dissertation]. Kyoto: Kyoto University,1973.
[8] Buhmann J.,Lades M.,Malsburg Von Der.Size and distortion invariant object recognition by hierarchical graph matching[A].Proceedings of IEEE Intl.Joint Conference on Neural Networks[C].San Diego,1990:411-416.
[9] Turk M and Pentland A.Eigenfaces for recognition[J].Cognitive Neuroscience,1991,3(1):71-86.
[10] Moghaddam B. , Penland A.. Probabilistic Visual Learning for Object Detection[A].The 5th International Conference on Computer Vision (ICCV’95) [C].Cambridge,MA,1995:786-793.
[11] Bellhumer P.N.,Hespanha J.,Kriegman D..Eigenfaces Vs.fisherfaces:Recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,Special Issue on Face Recognition,1997,l7(7):7l1-720.
[12] Bernhard Scholkopf, Alexander Smolla, and Klaus-Robert Muller, Nonlinear Component Analysis as a Kernel Eigenvalue Problem, Neural Computation,1998,10(5):1299.
[13]程永清,庄永明,汪华峰,杨静宇.一种有效的人脸识别方法,自动化学报,1993,19(1):54-61.
[14]周洪祥,何振亚,张辉.基于主元分析神经网络的人脸特征提取及识别研究.模式识别与人工智能,1996,9(1):52-57.
[15]王华,李介谷.人脸斜视图像的特征提取与恢复.上海交通大学学报,1997,31(1):101-104.
[16]张长水,彭辉等.基于K-L变换的人脸自动识别方法.清华大学学报(自然科学版本),1997,37(3):67-70.
[17]周激流等.人脸识别理论研究进展[J].计算机辅助设计与图形学学报,1999,11(2):180—184.
[18]贾云得,张小洵.基于互补子空间线性判别分析的人脸识别[J].北京理工大学学报,2006,26(3):206-210.
[19]高全学,潘泉,梁彦等.基于描述特征的人脸识别研究[J].自动化学报,2006,32(3):386-392.
[20]杨梅.基于LBP和混合2DPCA+PCA的人脸识别研究[D].南京:南京理工大学,2007.
[21] Kirby M and Sirovich L. Application of the KL procedure for the characterization of human faces. IEEE Trans. Pattern Anal. Machine Intell, 1990,12(1):103-108.
[22] Turk M and Pentland A, Face Recognition using eigenfaces, In Proceeding of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1994, pp:586-591.
[23] Pentland A et al. View-based modular eigenspaces for face recognition. Proc. IEEE Conf. on CVPR, 1994, 84-91.
[24] Swets D. and Weng J., Using Discriminant Eigenfeatures for Image Retrieval. IEEE Trans. PAMI, 1996,18(8):831-836.
[25] Yu H, Yang J, A Direct LDA Algorithm for High-Dimensional Data With Application to Face Recognition, Pattern Recognition, 2001, 34(10):2067-2070.
[26] Swets D, Weng J, Discriminant Analysis and Eigenspace Partition Tree for Face and Object Recognition from Views, in Proceedings of the International Conference on Automatic Face and Gesture Recognition, Killington, Vermont, USA, 1996, 192-197.
[27]刘永俊,陈才扣.基于差空间的最大散度差鉴别分析及人脸识别[J].计算机应用,2006,26(10):2460—2462.
[28]宋枫溪,程科,杨静宇.最大散度差和大间距线性投影与支持向量[J].自动化学报,2004,30(11):890—896.
[29]刘永俊,陈才扣.最大散度差鉴别分析及人脸识别[J].计算机工程与应用,2006,42(12):208—210.
[30] Liu C. J., Wechsler H., A shape and texture-based enhanced Fisher classifier for face recognition. IEEE Transactions on Image Processing, 2001, 10(4):598-608.
[31] Yang J., Yang J. Y., Generalized K-L transform based combined feature extraction. Pattern Recognition, 2002, 35(1): 295-297.
[32] Yang J., Yang J. Y., Zhang D., Lu J. F., Feature fusion: parallel strategy vs. serial strategy. Pattern Recognition, 2003, 36(6): 1369-1381.
[33] Hotelling H.Relations between two sets of variates.Biometrika,1936.8:321-377.
[34] Yang J., Yang J.Y., Generalized K-L transform based combined feature extraction. Pattern Recognition,2002,35(1):295-297.
[35]孙权森等,基于典型相关分析的组合特征抽取及脸像鉴别.计算机研究与发展,2005,42(4):614-621.
[36]权森,曾生根,王平安,夏德森.典型相关分析的理论及其在特征融合中的应用.计算机学报,2005,28(9):1524.1533.
[37]刘永俊.小样本问题下人脸图像的鉴别特征抽取方法研究[D].扬州:扬州大学,2007.
[2] Bertillon and J.Ruiz Solar. Pattern Recognition with Support Vector Machines volume 2388,1971,chapter On the Generalization of Kernel Machines,pages 24-39.Springer,August.
[3] Allen A L.Personal Descriptions,Butterworth,1950.
[4] Parker F I.Computer Generated Animation of Faces. In Processing ACM ann Conference, 1972, l:451-457.
[5] Goldation R J,Harmon L D,Lesk A B.Man manchine interaction in human face Identification. Bell Syst Tech Journal,1973,52:265-289.
[6] Kaya Y Kobayashi K.A Basic study on Human Recognition. In Frontiers of attern Recognition. New York:Academic,1971:265-289.
[7] Kanad T. Picture processing system by computer and recognition of human face [Ph D Dissertation]. Kyoto: Kyoto University,1973.
[8] Buhmann J.,Lades M.,Malsburg Von Der.Size and distortion invariant object recognition by hierarchical graph matching[A].Proceedings of IEEE Intl.Joint Conference on Neural Networks[C].San Diego,1990:411-416.
[9] Turk M and Pentland A.Eigenfaces for recognition[J].Cognitive Neuroscience,1991,3(1):71-86.
[10] Moghaddam B. , Penland A.. Probabilistic Visual Learning for Object Detection[A].The 5th International Conference on Computer Vision (ICCV’95) [C].Cambridge,MA,1995:786-793.
[11] Bellhumer P.N.,Hespanha J.,Kriegman D..Eigenfaces Vs.fisherfaces:Recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,Special Issue on Face Recognition,1997,l7(7):7l1-720.
[12] Bernhard Scholkopf, Alexander Smolla, and Klaus-Robert Muller, Nonlinear Component Analysis as a Kernel Eigenvalue Problem, Neural Computation,1998,10(5):1299.
[13]程永清,庄永明,汪华峰,杨静宇.一种有效的人脸识别方法,自动化学报,1993,19(1):54-61.
[14]周洪祥,何振亚,张辉.基于主元分析神经网络的人脸特征提取及识别研究.模式识别与人工智能,1996,9(1):52-57.
[15]王华,李介谷.人脸斜视图像的特征提取与恢复.上海交通大学学报,1997,31(1):101-104.
[16]张长水,彭辉等.基于K-L变换的人脸自动识别方法.清华大学学报(自然科学版本),1997,37(3):67-70.
[17]周激流等.人脸识别理论研究进展[J].计算机辅助设计与图形学学报,1999,11(2):180—184.
[18]贾云得,张小洵.基于互补子空间线性判别分析的人脸识别[J].北京理工大学学报,2006,26(3):206-210.
[19]高全学,潘泉,梁彦等.基于描述特征的人脸识别研究[J].自动化学报,2006,32(3):386-392.
[20]杨梅.基于LBP和混合2DPCA+PCA的人脸识别研究[D].南京:南京理工大学,2007.
[21] Kirby M and Sirovich L. Application of the KL procedure for the characterization of human faces. IEEE Trans. Pattern Anal. Machine Intell, 1990,12(1):103-108.
[22] Turk M and Pentland A, Face Recognition using eigenfaces, In Proceeding of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1994, pp:586-591.
[23] Pentland A et al. View-based modular eigenspaces for face recognition. Proc. IEEE Conf. on CVPR, 1994, 84-91.
[24] Swets D. and Weng J., Using Discriminant Eigenfeatures for Image Retrieval. IEEE Trans. PAMI, 1996,18(8):831-836.
[25] Yu H, Yang J, A Direct LDA Algorithm for High-Dimensional Data With Application to Face Recognition, Pattern Recognition, 2001, 34(10):2067-2070.
[26] Swets D, Weng J, Discriminant Analysis and Eigenspace Partition Tree for Face and Object Recognition from Views, in Proceedings of the International Conference on Automatic Face and Gesture Recognition, Killington, Vermont, USA, 1996, 192-197.
[27]刘永俊,陈才扣.基于差空间的最大散度差鉴别分析及人脸识别[J].计算机应用,2006,26(10):2460—2462.
[28]宋枫溪,程科,杨静宇.最大散度差和大间距线性投影与支持向量[J].自动化学报,2004,30(11):890—896.
[29]刘永俊,陈才扣.最大散度差鉴别分析及人脸识别[J].计算机工程与应用,2006,42(12):208—210.
[30] Liu C. J., Wechsler H., A shape and texture-based enhanced Fisher classifier for face recognition. IEEE Transactions on Image Processing, 2001, 10(4):598-608.
[31] Yang J., Yang J. Y., Generalized K-L transform based combined feature extraction. Pattern Recognition, 2002, 35(1): 295-297.
[32] Yang J., Yang J. Y., Zhang D., Lu J. F., Feature fusion: parallel strategy vs. serial strategy. Pattern Recognition, 2003, 36(6): 1369-1381.
[33] Hotelling H.Relations between two sets of variates.Biometrika,1936.8:321-377.
[34] Yang J., Yang J.Y., Generalized K-L transform based combined feature extraction. Pattern Recognition,2002,35(1):295-297.
[35]孙权森等,基于典型相关分析的组合特征抽取及脸像鉴别.计算机研究与发展,2005,42(4):614-621.
[36]权森,曾生根,王平安,夏德森.典型相关分析的理论及其在特征融合中的应用.计算机学报,2005,28(9):1524.1533.
[37]刘永俊.小样本问题下人脸图像的鉴别特征抽取方法研究[D].扬州:扬州大学,2007.