基于深度卷积稀疏自编码分层网络的人脸识别技术
|
摘要
面对海量人脸图像识别,传统特征提取方法难以提取有效特征,造成人脸识别准确率较低。提出了一种鲁棒的人脸特征提取算法,即利用深度卷积稀疏自编码网络自动学习人脸中丰富且识别力高的特征。该方法将卷积操作融入自编码网络中,同时加入稀疏化思想,从而形成深度卷积稀疏自编码分层网络(hierarchical deep convolution sparse autoencoder,HDCSAE);用该网络自动提取海量人脸图像的高层鲁棒特征,并将提取的特征作为SVM分类器的输入得到分类结果。在FERET人脸数据库下对该方法进行测试,识别率达到99.47%,比传统的基于提取人为定义特征的人脸识别方法的识别率有所提高。
In the face of massive face image recognition,traditional feature extraction methods are difficult to extract effective features,resulting in low face recognition accuracy.A robust face feature extraction algorithm is proposed,which uses the deep convolution sparse self-encoding network to automatically learn the features of the face that are rich and highly recognizable.This method integrates the convolution operation into the self-encoding network,and adds the sparse idea to form a deep convolution sparse autoencoder(HDCSAE);the network automatically extracts the high-level robust features of the massive face image,and uses the extracted features as the input of the SVM classifier to obtain the classification result.This method is tested under the FERET face database,and the recognition rate reaches 99.47%,which is better than that of the traditional face recognition method based on extracting artificially defined features.
In the face of massive face image recognition,traditional feature extraction methods are difficult to extract effective features,resulting in low face recognition accuracy.A robust face feature extraction algorithm is proposed,which uses the deep convolution sparse self-encoding network to automatically learn the features of the face that are rich and highly recognizable.This method integrates the convolution operation into the self-encoding network,and adds the sparse idea to form a deep convolution sparse autoencoder(HDCSAE);the network automatically extracts the high-level robust features of the massive face image,and uses the extracted features as the input of the SVM classifier to obtain the classification result.This method is tested under the FERET face database,and the recognition rate reaches 99.47%,which is better than that of the traditional face recognition method based on extracting artificially defined features.
引文
[1]BRUNELLI R,POGGIO T.Face recognition:features versus templates[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1993,15(10):1042-1052.
[2]WISKOTT L,FELLOUS J M,KUIGER N,et al.Face recognition by elastic bunch graph matching[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):775-779.
[3]CHEUNG K H,KONG A,YOU J,et al.A new approach to appearance-based face recognition[C]∥IEEE International Conference on Systems,Man and Cybernetics.2006:1686-1691.
[4]WRIGHT J,GANESH A,ZHOU Z,et al.Demo:robust face recognition via sparse representation[C]∥2008 8th IEEE International Conference on Automatic Face&Gesture Recognition.2008:1-2.
[5]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.
[6]WRIGHT J,YANG A Y,GANESH A,et al.Robust face recognition via sparse representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(2):210-227.
[7]杨荣根,任明武,杨静宇.基于稀疏表示的人脸识别方法[J].计算机科学,2010,37(9):267-269,278.YANG R G,REN M W,YANG J Y.Sparse representation based face recognition algorithm[J].Computer Science,2010,37(9):267-269,278.
[8]WANG W H,YANG J,XIAO J W,et al.Face recognition based on deep learning[C]∥International Conference on Human Centered Computing.2014:812-820.
[9]HU G,YANG Y,YI D,et al.When face recognition meets with deep learning:an evaluation of convolutional neural networks for face recognition[C]∥IEEE International Conference on Computer Vision Workshop.2016:384-392.
[10]ZHU Z,LUO P,WANG X,et al.Deep learning multi-view representation for face recognition[DB/OL].arXiv:1406.6947.[2017-03-16].https://arxiv.org.
[11]LIOU C Y,HUANG J C,YANG W C.Modeling word perception using the Elman network[J].Neurocomputing,2008,71(16/18):3150-3157.
[12]PIMENTEL C,DEMO-SOUZA R,UCHOA-FILHO B F,et al.Minimal trellis for systematic recursive convolutional encoders[C]∥IEEE International Symposium on Information Theory Proceedings.2011:2477-2481.
[13]HOYER P O,HYVRINEN A.A multi-layer sparse coding network learns contour coding from natural images[J].Vision Research,2002,42(12):1593-1605.
[14]WEN X,SHAO L,FANG W,et al.Efficient feature selection and classification for vehicle detection[J].IEEE Transactions on Circuits&Systems for Video Technology,2015,25(3):508-517.
[15]李辉,石波.基于卷积神经网络的人脸识别算法[J].软件导刊,2017,16(3):26-29.LI H,SHI B.Face recognition based on convolutional neural network[J].Software Guide,2017,16(3):26-29.
[16]冉鹏,王灵,李昕.一种基于栈式降噪自编码器的人脸识别方法[J].工业控制计算机,2016,29(9):100-101,104.RAN P,WANG L,LI X.A face recognition method based on stacked denoising auto-encoders[J].Industrial Control Computer,2016,29(9):100-101,104.
[17]张富铭.基于改进非负稀疏编码的人脸识别[D].武汉:武汉理工大学,2009.ZHANG F M.Face recognition based on improved non-negative sparse coding[D].Wuhan:Wuhan University of Technology,2009.
[2]WISKOTT L,FELLOUS J M,KUIGER N,et al.Face recognition by elastic bunch graph matching[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):775-779.
[3]CHEUNG K H,KONG A,YOU J,et al.A new approach to appearance-based face recognition[C]∥IEEE International Conference on Systems,Man and Cybernetics.2006:1686-1691.
[4]WRIGHT J,GANESH A,ZHOU Z,et al.Demo:robust face recognition via sparse representation[C]∥2008 8th IEEE International Conference on Automatic Face&Gesture Recognition.2008:1-2.
[5]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.
[6]WRIGHT J,YANG A Y,GANESH A,et al.Robust face recognition via sparse representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(2):210-227.
[7]杨荣根,任明武,杨静宇.基于稀疏表示的人脸识别方法[J].计算机科学,2010,37(9):267-269,278.YANG R G,REN M W,YANG J Y.Sparse representation based face recognition algorithm[J].Computer Science,2010,37(9):267-269,278.
[8]WANG W H,YANG J,XIAO J W,et al.Face recognition based on deep learning[C]∥International Conference on Human Centered Computing.2014:812-820.
[9]HU G,YANG Y,YI D,et al.When face recognition meets with deep learning:an evaluation of convolutional neural networks for face recognition[C]∥IEEE International Conference on Computer Vision Workshop.2016:384-392.
[10]ZHU Z,LUO P,WANG X,et al.Deep learning multi-view representation for face recognition[DB/OL].arXiv:1406.6947.[2017-03-16].https://arxiv.org.
[11]LIOU C Y,HUANG J C,YANG W C.Modeling word perception using the Elman network[J].Neurocomputing,2008,71(16/18):3150-3157.
[12]PIMENTEL C,DEMO-SOUZA R,UCHOA-FILHO B F,et al.Minimal trellis for systematic recursive convolutional encoders[C]∥IEEE International Symposium on Information Theory Proceedings.2011:2477-2481.
[13]HOYER P O,HYVRINEN A.A multi-layer sparse coding network learns contour coding from natural images[J].Vision Research,2002,42(12):1593-1605.
[14]WEN X,SHAO L,FANG W,et al.Efficient feature selection and classification for vehicle detection[J].IEEE Transactions on Circuits&Systems for Video Technology,2015,25(3):508-517.
[15]李辉,石波.基于卷积神经网络的人脸识别算法[J].软件导刊,2017,16(3):26-29.LI H,SHI B.Face recognition based on convolutional neural network[J].Software Guide,2017,16(3):26-29.
[16]冉鹏,王灵,李昕.一种基于栈式降噪自编码器的人脸识别方法[J].工业控制计算机,2016,29(9):100-101,104.RAN P,WANG L,LI X.A face recognition method based on stacked denoising auto-encoders[J].Industrial Control Computer,2016,29(9):100-101,104.
[17]张富铭.基于改进非负稀疏编码的人脸识别[D].武汉:武汉理工大学,2009.ZHANG F M.Face recognition based on improved non-negative sparse coding[D].Wuhan:Wuhan University of Technology,2009.