改进的二维主成分分析的人脸识别新算法
|
摘要
传统二维主成分分析(2DPCA)中的变换只提取人脸图像数据的行内特征,特征提取的方向相对比较单一,没有考虑到其他方向上的特征提取。为了多角度提取图像的特征,识别提供更丰富的信息,文中提出一种改进的2DPCA人脸识别算法。该算法先将人脸图像进行倾斜角度自矫正,同时提取图片的低频信息,再利用改进的感知哈希技术得到图像的"指纹",然后将自矫正后的人脸图片进行多角度旋转,并分别提取特征,得到多角度旋转后的图像特征信息。最后将新算法在ORL
The traditional two-dimensional principal component analysis(2DPCA)only extracts the in-line features of hu-man face image data during the transformation,and the feature extraction is relatively single direction-oriented,without consider-ing feature extractions in other directions. Therefore,a human face recognition algorithm based on the improved 2DPCA is pro-posed,so as to extract image features from multiple angles and provide more abundant information for recognition. In the algo-rithm,the tilt angle self-correction is conducted for human face images,and meanwhile,the low frequency information of imag-es is extracted. The improved perceptual hash technology is used to obtain the "fingerprints" of images. The multi-angle rotation is conducted for human face images after self-correction. Features are respectively extracted to obtain the image feature informa-tion after multi-angle rotation. The new algorithm was tested with the ORL human face library. The results show that the im-proved algorithm is superior to the traditional 2DPCA.
The traditional two-dimensional principal component analysis(2DPCA)only extracts the in-line features of hu-man face image data during the transformation,and the feature extraction is relatively single direction-oriented,without consider-ing feature extractions in other directions. Therefore,a human face recognition algorithm based on the improved 2DPCA is pro-posed,so as to extract image features from multiple angles and provide more abundant information for recognition. In the algo-rithm,the tilt angle self-correction is conducted for human face images,and meanwhile,the low frequency information of imag-es is extracted. The improved perceptual hash technology is used to obtain the "fingerprints" of images. The multi-angle rotation is conducted for human face images after self-correction. Features are respectively extracted to obtain the image feature informa-tion after multi-angle rotation. The new algorithm was tested with the ORL human face library. The results show that the im-proved algorithm is superior to the traditional 2DPCA.
引文
[1]CHOI K,TOH K A,BYUN H.Incremental face recognition for large-scale social network services[J].Pattern recognition,2012,45(8):2868-2883.
[2]MASHHOORI A,JAHROMI M Z.Block-wise two-directional2DPCA with ensemble learning for face recognition[J].Neurocomputing,2013,108:111-117.
[3]黄慧,路翀.基于PCA与2DPCA的少数民族人脸识别比较[J].价值工程,2016,35(11):223-224.HUANG Hui,LU Chong.Ethnic minorities face recognition based on PCA and 2DPCA[J].Value engineering,2016,35(11):223-224.
[4]冯飞,姜宝华,刘培学,等.改进2DPCA算法在人脸识别中的应用[J].计算机科学,2017,44(z2):267-268.FENG Fei,JIANG Baohua,LIU Peixue,et al.Application of improved 2DPCA algorithm in face recognition[J].Computer science,2017,44(S2):267-268.
[5]LU G,ZOU J,WANG Y.Incremental complete LDA for face recognition[J].Pattern recognition,2012,45(7):2510-2521.
[6]杨梅芳,石义龙.基于2DPCA+PCA与SVM的人脸识别[J].信息技术,2018(2):32-36.YANG Meifang,SHI Yilong.Face recognition based on 2DP-CA+PCA and SVM[J].Information technology,2018(2):32-36.
[7]YANG J,ZHANG D,FRANGI A F,et al.Two-dimensional PCA:a new approach to appearance-based face representation and recognition[J].IEEE transactions on pattern analysis and machine intelligence,2004,26(1):131-137.
[8]黄海波.基于2DPCA和PCA的特征提取的人脸识别算法研究[D].昆明:昆明理工大学,2014.HUANG Haibo.Research on human face recognition algorithm using 2DPCA and PCA based feature extraction[D].Kunming:Kunming University of Science and Technology,2014.
[9]李娟,何伟,张玲,等.双向压缩的2DPCA与PCA相结合的人脸识别算法[J].计算机应用,2009,29(z1):245-246.LI Juan,HE Wei,ZHANG Ling,et al.Face recognition combined bidirectional 2DPCA with PCA[J].Journal of computer applications,2009,29(S1):245-246.
[10]王心醉,李岩,郭立红,等.基于双向PCA和K近邻的人脸识别算法[J].解放军理工大学学报(自然科学版),2010,11(6):623-627.WANG Xinzui,LI Yan,GUO Lihong,et al.Face recognition algorithm based on BD-PCA and KNN[J].Journal of PLAUniversity of Science and Technology(Natural science edition),2010,11(6):623-627.
[11]李童.基于行列双向优化的2DPCA人脸识别方法研究[D].重庆:重庆师范大学,2013.LI Tong.Research on face recognition using bidirectional optimization of 2DPCA method[D].Chongqing:Chongqing Normal University,2013.
[12]柳春,杨绍华,方强,等.基于对角PCA的人脸识别方法[J].电脑与信息技术,2012,20(4):20-21.LIU Chun,YANG Shaohua,FANG Qiang,et al.A face recognition method based on diagonal-PCA[J].Computer and information technology,2012,20(4):20-21.
[13]张杨,张仁杰.基于改进PCA算法的人脸识别[J].软件导刊,2018,17(1):32-34.ZHANG Yang,ZHANG Renjie.Face recognition based on improved PCA algorithm[J].Software guide,2018,17(1):32-34.
[14]孙艳娜.基于2DPCA的人脸识别方法[D].西安:西安电子科技大学,2013.SUN Yanna.Face recognition based on 2DPCA[D].Xi’an:Xidian University,2013.64
[15]LI T,LI M,GAO Q,et al.F-norm distance metric based robust 2DPCA and face recognition[J].Neural networks,2017,94:204-211.
[16]刘兆庆.图像感知哈希若干关键技术研究[D].哈尔滨:哈尔滨工业大学,2013.LIU Zhaoqing.Research on key techniques of image perceptual hashing[D].Harbin:Harbin Institute of Technology,2013.
[17]谭子尤,梁靖.基于PCA+2DPCA的人脸识别方法分析[J].吉首大学学报(自然科学版),2011,32(3):55-58.TAN Ziyou,LIANG Jing.Face recognition based on linear transformation theory[J].Journal of Jishou University(Natural science edition),2011,32(3):55-58.
[2]MASHHOORI A,JAHROMI M Z.Block-wise two-directional2DPCA with ensemble learning for face recognition[J].Neurocomputing,2013,108:111-117.
[3]黄慧,路翀.基于PCA与2DPCA的少数民族人脸识别比较[J].价值工程,2016,35(11):223-224.HUANG Hui,LU Chong.Ethnic minorities face recognition based on PCA and 2DPCA[J].Value engineering,2016,35(11):223-224.
[4]冯飞,姜宝华,刘培学,等.改进2DPCA算法在人脸识别中的应用[J].计算机科学,2017,44(z2):267-268.FENG Fei,JIANG Baohua,LIU Peixue,et al.Application of improved 2DPCA algorithm in face recognition[J].Computer science,2017,44(S2):267-268.
[5]LU G,ZOU J,WANG Y.Incremental complete LDA for face recognition[J].Pattern recognition,2012,45(7):2510-2521.
[6]杨梅芳,石义龙.基于2DPCA+PCA与SVM的人脸识别[J].信息技术,2018(2):32-36.YANG Meifang,SHI Yilong.Face recognition based on 2DP-CA+PCA and SVM[J].Information technology,2018(2):32-36.
[7]YANG J,ZHANG D,FRANGI A F,et al.Two-dimensional PCA:a new approach to appearance-based face representation and recognition[J].IEEE transactions on pattern analysis and machine intelligence,2004,26(1):131-137.
[8]黄海波.基于2DPCA和PCA的特征提取的人脸识别算法研究[D].昆明:昆明理工大学,2014.HUANG Haibo.Research on human face recognition algorithm using 2DPCA and PCA based feature extraction[D].Kunming:Kunming University of Science and Technology,2014.
[9]李娟,何伟,张玲,等.双向压缩的2DPCA与PCA相结合的人脸识别算法[J].计算机应用,2009,29(z1):245-246.LI Juan,HE Wei,ZHANG Ling,et al.Face recognition combined bidirectional 2DPCA with PCA[J].Journal of computer applications,2009,29(S1):245-246.
[10]王心醉,李岩,郭立红,等.基于双向PCA和K近邻的人脸识别算法[J].解放军理工大学学报(自然科学版),2010,11(6):623-627.WANG Xinzui,LI Yan,GUO Lihong,et al.Face recognition algorithm based on BD-PCA and KNN[J].Journal of PLAUniversity of Science and Technology(Natural science edition),2010,11(6):623-627.
[11]李童.基于行列双向优化的2DPCA人脸识别方法研究[D].重庆:重庆师范大学,2013.LI Tong.Research on face recognition using bidirectional optimization of 2DPCA method[D].Chongqing:Chongqing Normal University,2013.
[12]柳春,杨绍华,方强,等.基于对角PCA的人脸识别方法[J].电脑与信息技术,2012,20(4):20-21.LIU Chun,YANG Shaohua,FANG Qiang,et al.A face recognition method based on diagonal-PCA[J].Computer and information technology,2012,20(4):20-21.
[13]张杨,张仁杰.基于改进PCA算法的人脸识别[J].软件导刊,2018,17(1):32-34.ZHANG Yang,ZHANG Renjie.Face recognition based on improved PCA algorithm[J].Software guide,2018,17(1):32-34.
[14]孙艳娜.基于2DPCA的人脸识别方法[D].西安:西安电子科技大学,2013.SUN Yanna.Face recognition based on 2DPCA[D].Xi’an:Xidian University,2013.64
[15]LI T,LI M,GAO Q,et al.F-norm distance metric based robust 2DPCA and face recognition[J].Neural networks,2017,94:204-211.
[16]刘兆庆.图像感知哈希若干关键技术研究[D].哈尔滨:哈尔滨工业大学,2013.LIU Zhaoqing.Research on key techniques of image perceptual hashing[D].Harbin:Harbin Institute of Technology,2013.
[17]谭子尤,梁靖.基于PCA+2DPCA的人脸识别方法分析[J].吉首大学学报(自然科学版),2011,32(3):55-58.TAN Ziyou,LIANG Jing.Face recognition based on linear transformation theory[J].Journal of Jishou University(Natural science edition),2011,32(3):55-58.