一种基于相关性投影的人脸性别分类算法
|
摘要
针对于人脸图像检测的有效利用性,为了提高其检测的性能,提出一种新的基于监督学习的优化相关性投影(ORP)人脸性别分类算法,并将其应用到基于Eigenface算法与Fisherface算法的人脸识别中,以及应用WPCA到基于PGA的性别分类中。本文算法首先基于带权主成分分析(WPCA)算法来降低脸部维度,将脸部特征提取出;然后,对其进行优化,同时计算ORP的误差函数;最后,最小化脸部ORP误差函数,计算特征向量的欧式距离,进行人脸性别分类。将提出方法与传统方法进行对比,在FERET数据库上进行了实验,证明了本文方法的有效性,获得了优于传统方法的识别率。
According to the effective use of face image detection,in order to improve the performance of the existing face detection algorithms,a new method based on supervised learning for face gender classification of relevance maps is proposed.This method can be applied to different facial analysis tasks,and has been successfully applied to face recognition based on Eigenface algorithm and Fisherface algorithm,also applied to the principal component analysis(PGA)in the gender classification.Firstly,the algorithm reduces the dimension of the face based on the weighted PCA algorithm,and facial features are extracted.Then,the algorithm is optimized and the error function of correlation projection is calculated.Finally,the correlation error function in facial projection is minimized and the Euclidean distance of the feature vector is calculated for face gender classification.The proposed method is compared with other feature extraction methods,and the corresponding experiments are carried out on the FERET database.The results prove the method is efficient and can obtain higher recognition accuracy than traditional methods.
According to the effective use of face image detection,in order to improve the performance of the existing face detection algorithms,a new method based on supervised learning for face gender classification of relevance maps is proposed.This method can be applied to different facial analysis tasks,and has been successfully applied to face recognition based on Eigenface algorithm and Fisherface algorithm,also applied to the principal component analysis(PGA)in the gender classification.Firstly,the algorithm reduces the dimension of the face based on the weighted PCA algorithm,and facial features are extracted.Then,the algorithm is optimized and the error function of correlation projection is calculated.Finally,the correlation error function in facial projection is minimized and the Euclidean distance of the feature vector is calculated for face gender classification.The proposed method is compared with other feature extraction methods,and the corresponding experiments are carried out on the FERET database.The results prove the method is efficient and can obtain higher recognition accuracy than traditional methods.
引文
[1]Peter N Belhumeur,Jo珘ao P Hespanha,David J Kriegman.Eigenfaces vs.fisherfaces:recognition using class specific linear projection[J].IEEE Trans.Pattern Analysis and Machine Intelligence,2012,19(7):711-720.
[2]Liu X N.Research on face recognition technology based on 3D model[D].Doctoral Dissertation of Northwestern University,Journal of Shaanxi University of Science&Technology,2011.
[3]Buchala S,Davey N,Gale T M,et al.Principal component analysis of gender,ethnicity,age,and identity of face images[A].Proc.of IEEE ICMI[C].2015,118-122.
[4]Jain A,Huang J,Integrating independent components and linear discriminant analysis for gender classification[A].Proc.of FGR2014[C].2014,159-163.
[5]Moghaddam B,Yang M H.Learning gender with support faces[J].IEEE Transactions on Pattern Analysis and Machine Intelligence[C].2012,24(5):707-711.
[6]CUI Peng,ZHANG Xue-ting.A block-based bi-directional Fisher linear discriminant analysis on face recognition[J].Journal of Optoelectronics·Laser,2016,27(4):421-428.崔鹏,张雪婷.基于块双向Fisher线性判别分析人脸识别[J].光电子·激光,2016,27(4):421-428.
[7]Lades M,Buhmann J,Lades M,et al.Distortion invariant object detection in the dynamic link architecture[J].IEEE Trans on Computer,2013,42(3):300-310.
[8]Wiskott L,Basuki,Andy,et al.Face detection by elastic bunch graph matching[A].Proc.of Intelligent Biometric Techniques in Fingerprint and Face Detection[C].2014,355-396.
[9]Saatci Y,Town C P.Cascaded classification of gender and facial expression using active appearance models[A].Proc.of International Conference on Automatic Face and Gesture Recognition[C].2015,136-142.
[10]Brunelli R,Poggio T.HyberBF networks for gender classification[A].Proc.of DARPA Image Understanding Workshop[C].2012,311-314.
[11]Cottrell G W,Metcalfe J.Face,emotion,and gender recognition using holons[A].Advances in Neural Information Processing Systems 3[C].2011,564-571.
[12]Turk M,Pentland A.Face recognition using eigenfaces[A].Proc.of IEEE Conference on Computer Vision and Pattern Recognition[C].1991,586-591.
[13]Sun Z,Bebis G,Yuan X,et al.Genetic feature subset selection for gender classification:A comparison study[A].Proc.of Well Actually Computer Vision(WACV)[C].2012,165-170.
[14]Yambor W,Draper B,Beveridge R.Analyzing PCA-based face recognition algorithms:eigenvector selection and distance measures[A].Proc.of Empirical Evaluation Methods in Computer Vision[C].2002,85-94.
[15]Pennec X.Probabilities and statistics on riemannian manifolds:ageometric approach[R].Technical Report RR-5093,HAL-INRIA,2004.
[16]WU Di,TANG Yong-qi,LIN Guo-han.Roboust face recognition based on significance local directional pattern and deep learning[J].Journal of Optoelectronics·Laser,2016,27(6):655-661.吴迪,唐勇奇,林国汉.基于显著性局部定向模式和深度学习的鲁棒人脸识别[J].光电子·激光,2016,27(6):655-661.
[17]CUI Peng,ZHANG Ru-bo.Face feature extraction method based on part of labeled data[J].Journal of Optoelectronics·Laser,2012,23(3):554-560.崔鹏,张汝波.基于部分标记数据进行人脸图像特征提取[J].光电子·激光,2012,23(3):554-560.
[18]Wu J,Smith W A P,Hancock E R.Weighted principal geodesic analysis for facial gender classification[A].Proc.of British Machine Vision Conference(BMVC)[C].2007,331-339.
[19]Kawulok M.Optimization of face relevance maps with total classification error minimization[A].Proc.of Lecture Notes in Computer Science[C].2008,5112:935-944.
[20]Phillips P J,Moon H,Rizvi S A,et al.The FERET evaluation methodology for face-recognition algorithms[J].IEEE Trans,Pattern Analysis and Machine Intelligence,2010,22(10):1090-1104,2010.
[2]Liu X N.Research on face recognition technology based on 3D model[D].Doctoral Dissertation of Northwestern University,Journal of Shaanxi University of Science&Technology,2011.
[3]Buchala S,Davey N,Gale T M,et al.Principal component analysis of gender,ethnicity,age,and identity of face images[A].Proc.of IEEE ICMI[C].2015,118-122.
[4]Jain A,Huang J,Integrating independent components and linear discriminant analysis for gender classification[A].Proc.of FGR2014[C].2014,159-163.
[5]Moghaddam B,Yang M H.Learning gender with support faces[J].IEEE Transactions on Pattern Analysis and Machine Intelligence[C].2012,24(5):707-711.
[6]CUI Peng,ZHANG Xue-ting.A block-based bi-directional Fisher linear discriminant analysis on face recognition[J].Journal of Optoelectronics·Laser,2016,27(4):421-428.崔鹏,张雪婷.基于块双向Fisher线性判别分析人脸识别[J].光电子·激光,2016,27(4):421-428.
[7]Lades M,Buhmann J,Lades M,et al.Distortion invariant object detection in the dynamic link architecture[J].IEEE Trans on Computer,2013,42(3):300-310.
[8]Wiskott L,Basuki,Andy,et al.Face detection by elastic bunch graph matching[A].Proc.of Intelligent Biometric Techniques in Fingerprint and Face Detection[C].2014,355-396.
[9]Saatci Y,Town C P.Cascaded classification of gender and facial expression using active appearance models[A].Proc.of International Conference on Automatic Face and Gesture Recognition[C].2015,136-142.
[10]Brunelli R,Poggio T.HyberBF networks for gender classification[A].Proc.of DARPA Image Understanding Workshop[C].2012,311-314.
[11]Cottrell G W,Metcalfe J.Face,emotion,and gender recognition using holons[A].Advances in Neural Information Processing Systems 3[C].2011,564-571.
[12]Turk M,Pentland A.Face recognition using eigenfaces[A].Proc.of IEEE Conference on Computer Vision and Pattern Recognition[C].1991,586-591.
[13]Sun Z,Bebis G,Yuan X,et al.Genetic feature subset selection for gender classification:A comparison study[A].Proc.of Well Actually Computer Vision(WACV)[C].2012,165-170.
[14]Yambor W,Draper B,Beveridge R.Analyzing PCA-based face recognition algorithms:eigenvector selection and distance measures[A].Proc.of Empirical Evaluation Methods in Computer Vision[C].2002,85-94.
[15]Pennec X.Probabilities and statistics on riemannian manifolds:ageometric approach[R].Technical Report RR-5093,HAL-INRIA,2004.
[16]WU Di,TANG Yong-qi,LIN Guo-han.Roboust face recognition based on significance local directional pattern and deep learning[J].Journal of Optoelectronics·Laser,2016,27(6):655-661.吴迪,唐勇奇,林国汉.基于显著性局部定向模式和深度学习的鲁棒人脸识别[J].光电子·激光,2016,27(6):655-661.
[17]CUI Peng,ZHANG Ru-bo.Face feature extraction method based on part of labeled data[J].Journal of Optoelectronics·Laser,2012,23(3):554-560.崔鹏,张汝波.基于部分标记数据进行人脸图像特征提取[J].光电子·激光,2012,23(3):554-560.
[18]Wu J,Smith W A P,Hancock E R.Weighted principal geodesic analysis for facial gender classification[A].Proc.of British Machine Vision Conference(BMVC)[C].2007,331-339.
[19]Kawulok M.Optimization of face relevance maps with total classification error minimization[A].Proc.of Lecture Notes in Computer Science[C].2008,5112:935-944.
[20]Phillips P J,Moon H,Rizvi S A,et al.The FERET evaluation methodology for face-recognition algorithms[J].IEEE Trans,Pattern Analysis and Machine Intelligence,2010,22(10):1090-1104,2010.