多层融合深度局部PCA子空间稀疏优化特征提取模型

doi:10.3969/j.issn.0372-2112.2017.10.011

摘要/Abstract

摘要： 子空间方法是主要利用全局信息的经典模式识别方法，随着深度学习思想的引入，局部自学习结构特征模型得到大家的关注.利用深度学习原理，本文提出一种多层融合的深度局部子空间稀疏优化特征自学习抽取模型解决目标识别问题.首先，对训练样本集通过最小化重构误差得到第一层的主成分（Principal Component Analysis，PCA）特征映射矩阵；然后，通过L₁范数约束对特征映射结果进行稀疏优化，提高算法鲁棒性.接着，在第二层映射层以第一层的特征输出为输入，进行同样的特征矩阵学习操作，最终将图像映射至深层PCA子空间；然后，对各个映射层的特征提取结果进行加权融合，进行二值化哈希编码和直方图分块编码，提取图像的深度子空间稀疏特征.在FERET、AR、Yale等经典人脸数据库以及MNIST、CIFAR-10等目标数据库上的实验结果表明，该算法可以取得较高的识别率以及较好的光照、表情、人脸朝向鲁棒性，并且相对于卷积神经网络等深度学习框架具有结构简洁、收敛速度快等优点.

关键词: 深度学习, 多层融合, 子空间, 稀疏优化

Abstract: Subspace method is classical pattern recognition method,that uses global information mainly to denote an image.Recently,with the introduction of deep learning,the feature extraction model based on local self-learning has attracted more and more attention.By using the theory of deep learning,this paper presents a new feature extraction model based on multi-layered deep local subspace sparse optimization to solve the problem of object recognition.Firstly,we calculate the PCA mapping matrix on the first layer by minimizing the reconstruction error on the training sample set,then we optimize the feature mapping results through L₁ norm to enhance the robustness of algorithm.Secondly,we use the output of the first layer as the input of second layer,then we implement same actions of feature learning.In this way we can map the image to deep PCA subspace.Finally we merge these feature extraction results from different layers with weighting and encode the merged feature with binary hash code and histogram segment code.After that,we obtain the multi-layered deep local subspace sparse feature.The experimental results on face database of FERET、AR、Yale and target database of MNIST、CIFAR-10 show that this feature extraction model can achieve high recognition rate and robustness for illumination,expression and pose.At the same time,compared with the convolutional neural networks,our algorithm owns the advantages of simple structure and fast convergent rate.

Key words: deep learning, multi-layered fusion, subspace, sparse optimization

中图分类号:

TP391.4

胡正平, 陈俊岭. 多层融合深度局部PCA子空间稀疏优化特征提取模型[J]. 电子学报, 2017, 45(10): 2383-2389.

HU Zheng-ping, CHEN Jun-ling. Feature Extraction Model Based on Multi-layered Deep Local Subspace Sparse Optimization[J]. Acta Electronica Sinica, 2017, 45(10): 2383-2389.

参考文献

[1] TURK M,PENTLAND A.Eigenfaces for recognition[J].Journal of Cognitive Neuroscience,1991,3(1):71-86.
[2] BELHUMEUR P N,HESPANHA J P,KRIEGMAN D.Eigenfaces vs Fisherfaces:recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):711-720.
[3] LI Yong-qiang,PAN Jin.Face recognition algorithm based on improved BP neural network[J].International Journal of Security and Its Applications,2015,9(5):175-184.
[4] 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.
[5] HINTON G,SALAKHUTDINOV R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
[6] CHAN Tsung Han,JIA Kui,GAO Shenghua,LU Jiwen,ZENG Zinan,MA Yi.PCANet:A simple deep learning baseline for image classification[J].IEEE Transactions on Image Processing,2015,24(12):5017-5032.
[7] TAN Hengliang,YANG Bing,MA Zhengming.Face recognition based on the fusion of global and local HOG features of face images[J].Computer Vision,2014,8(3):224-234.
[8] LIU Chengjun,WECHSLER H.Independent component analysis of Gabor features for face recognition[J].IEEE Transactions on Neural Networks,2003,14(4):919-928.
[9] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[A].Advances in Neural Information Processing Systems[C].Lake Tahoe,NV,United states:NIPS,2012.1097-1105.
[10] TAIGMAN Y,YANG Ming,RANZATO M,WOLF L.DeepFace:Closing the gap to human-level performance in face verification[A].IEEE Conference on Computer Vision and Pattern Recognition (CVPR)[C].Columbus,OH,United states:CVPR,2014.1701-1708.
[11] SUN Yi,WANG Xiaogang,TANG Xiaoou.Deep learning face representation from predicting 10,000 classes[A].IEEE Conference on Computer Vision and Pattern Recognition (CVPR)[C].Columbus,OH,United states:CVPR,2014.1891-1898.
[12] SUN Y,CHEN Y,WANG X,et al.Deep learning face representation by joint identification-verification[A].Advances in Neural Information Processing Systems[C].Montreal,Canada:NIPS,2014.1988-1996.
[13] SUN Yi,WANG Xiaogang,TANG Xiaoou.Deeply learned face representations are sparse,selective,and robust[A].The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)[C].Boston,MA,USA:CVPR,2015.2892-2900.
[14] LU Jiwen,LIONG V E,WANG Gang,MOULIN P.Joint feature learning for face recognition[J].IEEE Transactions on Information Forensics and Security,2015,10(7):1371-1383.
[15] LIONG V E,LU J,WANG G.Face recognition using Deep PCA[A].IEEE Transactions on Information,Communications and Signal Processing (ICICS)[C].Beijing,China:ICICS,2013.1-5.
[16] CHAN T H,JIA K,GAO S,et al.PCANet:A simple deep learning baseline for image classification?[J].IEEE Transactions on Image Processing,2015,24(12):5017-5032.
[17] 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.
[18] HUANG J,YUAN C.Weighted-PCANet for Face Recognition[A].Neural Information Processing[C].Montréal,Quebec,Canada:Springer International Publishing,2015.246-254.
[19] 焦李成,赵进,杨淑媛,刘芳,谢雯.稀疏认知学习、计算与识别的研究进展[J].计算机学报,2015,38(10):1-18. JIAO Li-cheng,ZHAO Jin,YANG Shu-yuan,LIU Fang,XIE Wen.Research advances on sparse cognitive learning,computing and recognition[J].Chinese Journal of Computers,2015,38(10):1-18.(in Chinese)
[20] WANG Xiaolong,GUO Rui,KAMBHAMETTU C.Deeply-learned feature for age estimation[A].IEEE Winter Conference on Applications of Computer Vision (WACV)[C].Waikoloa,HI,United states:IEEE,2015.534-541.
[21] NGIAM Jiquan,KOH Pangwei,Chen Zhenghao,BHASKAR Sonia,NG ANDREW Y.Sparse filtering[A].Advances in Neural Information Processing Systems[C].Granada,Spain:NIPS,2011.1125-1133.
[22] DAHL G E,SAINATH T N,HINTON G E.Improving deep neural networks for LVCSR using rectified linear units and dropout[A].IEEE International Conference on Acoustics,Speech and Signal Processing (ICASSP)[C].Vancouver,BC,Canada:IEEE,2013.8609-8613.
[23] WANG D,LU H.Object tracking via 2DPCA and L1-regularization[J].Signal Processing Letters,2012,19(11):711-714.