[1]时中荣,王 胜,刘传才.基于L2,p矩阵范数稀疏表示的图像分类方法[J].南京理工大学学报(自然科学版),2017,41(01):80.[doi:10.14177/j.cnki.32-1397n.2017.41.01.011]
 Shi Zhongrong,Wang Sheng,Liu Chuancai.Sparse representation via L2,p norm for image classification[J].Journal of Nanjing University of Science and Technology,2017,41(01):80.[doi:10.14177/j.cnki.32-1397n.2017.41.01.011]
点击复制

基于L2,p矩阵范数稀疏表示的图像分类方法()
分享到:

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

卷:
41卷
期数:
2017年01期
页码:
80
栏目:
出版日期:
2017-02-28

文章信息/Info

Title:
Sparse representation via L2,p norm for image classification
文章编号:
1005-9830(2017)01-0080-10
作者:
时中荣1王 胜2刘传才1
1.南京理工大学 计算机科学与工程学院,江苏 南京 210094; 2.河南大学 图像处理与模式识别研究所,河南 开封 475004
Author(s):
Shi Zhongrong1Wang Sheng2Liu Chuancai1
1.School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China; 2.Institute of Image Processing and Pattern Recognition,Henan University,Kaifeng 475004,China
关键词:
图像分类 稀疏表示 稀疏分类 矩阵范数 稀疏编码 字典学习 稀疏正则项 稀疏诱导范数
Keywords:
image classification sparse representation sparse classification matrix norm sparse coding dictionary learning sparse regularization sparse-inducing norm
分类号:
TP391.41
DOI:
10.14177/j.cnki.32-1397n.2017.41.01.011
摘要:
为了提高基于稀疏表示分类算法的分类精度,该文充分利用同类样本的非零系数高度集中的特点,提出一种用l2,p矩阵范数进行稀疏约束的基于稀疏表示的分类方法。该算法的训练阶段,构造的目标函数主要包括三个部分:重构误差、稀疏矩阵类内一致性约束、稀疏矩阵类间不一致性约束,其中的稀疏矩阵类内一致性约束用l2,p矩阵范数实现。该算法的测试阶段,计算新样本的稀疏重构系数以用于分类。和传统的基于稀疏表示的分类方法比较,该方法求稀疏重构系数时对样本不再单个处理,而是对同类样本整体处理,且充分利用同类样本的相似性和不同类样本的相异性,提高了基于稀疏表示的图像分类方法的分类精度。实验结果表明:该方法进一步提高了图像分类的准确率,在AR、Extended Yale B和Fifteen Scene Category数据库上和基于稀疏表示的分类方法(Sparse representation based classification,SRC)相比较,识别率分别提高了20.11%、20.88%和2.13%。
Abstract:
For the sparse representation-based classification method,since the non-zero elements of sparse coefficients with the same class are concentrated in a few rows,we propose to regularize the coefficient matrix using an l2,p matrix norm.In the training phase of the algorithm,the objective function consists of three parts:reconstruction error,sparse regularization,and inconsistency of reconstruction coefficients between different classes.The sparse regularization term is implemented by an l2,p matrix norm.In the test phase,the sparse reconstruction coefficient of a new sample is found using the dictionary learned in the training phase.Finally,the new sample is classified according to the sparse reconstruction coefficient.Compared with the traditional classification method based on sparse representation,the proposed method does not process a single sample to find its sparse reconstruction coefficient,but the whole sample matrix can be processed,this takes full advantage of the similarity among the same class.The experimental results show that this method can improve the accuracies of image classification 20.11%,20.88%,and 2.13% compared with a baseline SRC(Sparse representation based classification)method in AR,Extended Yale B,and Fifteen Scene Category databases,respectively.This method makes full use of the similarity of the same class and improves the accuracy of the image classification based on sparse representation.

参考文献/References:

[1] 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.
[2]Poppe R.A survey on vision-based human action recognition[J].Image and Vision Computing,2010,28(6):976-990.
[3]Datta R,Joshi D,Li J,et al.Image retrieval:ideas,influences,and trends of the new age[J].ACM Computing Surveys,2008,40(2):5.
[4]刘梓,宋晓宁,於东军,等.基于多成分字典和稀疏表示的超分辨率重建算法[J].南京理工大学学报:自然科学版,2014,38(1):1-5.
Liu Zi,Song Xiaoning,Yu Dongjun,et al.Super-resolution reconstruction algorithm based on multi-component dictionary and sparse representation[J].Journal of Nanjing University of Science and Technology,2014,38(1):1-5.
[5]Beaulieu-Jones B K,Greene C S.Semi-supervised learning of the electronic health record with denoising autoencoders for phenotype stratification[J].bioRxiv:039800,2016.
[6]Mairal J,Bach F,Ponce J.Task-driven dictionary learning[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(4):791-804.
[7]Aharon M,Elad M,Bruckstein A.K-SVD:An algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Transactions on Signal Processing,2006,54(11):4311-4322.
[8]Zhang Qiang,Li Baoxin.Discriminative K-SVD for dictionary learning in face recognition[C]//2010 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).San Francisco,USA:IEEE,2010:2691-2698.
[9]Jiang Zhuolin,Lin Zhe,Davis L S.Label consistent K-SVD:Learning a discriminative dictionary for recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(11):2651-2664.
[10]Yang Meng,Zhang Lei,Feng Xiangchu,et al.Sparse representation based fisher discrimination dictionary learning for image classification[J].International Journal of Computer Vision,2014,109(3):209-232.
[11]Yang Jianchao,Yu Kai,Gong Yihong,et al.Linear spatial pyramid matching using sparse coding for image classification[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Miami,USA:IEEE,2009:1794-1801.
[12]Do T H,Tabbone S,Terrades O R.Sparse representation over learned dictionary for symbol recognition[J].Signal Processing,2016,125:36-47.
[13]Lu Na,Miao Hongyu.Structure constrained nonnegative matrix factorization for pattern clustering and classification[J].Neurocomputing,2016,171:400-411.
[14]Lu Zhenyu,Zhang Linghua.Face recognition algorithm based on discriminative dictionary learning and sparse representation[J].Neurocomputing,2016,174:749-755.
[15]Wu Fei,Jing Xiaoyuan,You Xinge,et al.Multi-view low-rank dictionary learning for image classification[J].Pattern Recognition,2016,50:143-154.
[16]Yuan Li,Liu Wei,Li Yang.Non-negative dictionary based sparse representation classification for ear recognition with occlusion[J].Neurocomputing,2016,171:540-550.
[17]宋晓宁,薛益时,於东军,等.一种优化恢复压缩传感矩阵的模糊自适应方法[J].南京理工大学学报,2013,37(4):479-485.
Song Xiaoning,Xue Yishi,Yu Dongjun,et al.Fuzzy adaptive method for optimizing recovery of compressive sensing matrix[J].Journal of Nanjing University of Science and Technology,2013,37(4):479-485.
[18]张少辉,王迤冉.用于图像识别的稀疏高斯编码[J].南京理工大学学报,2016,40(1):61-66.
Zhang Shaohui,Wang Yiran.Sparse Gaussian coding for image recognition[J].Journal of Nanjing University of Science and Technology,2016,40(1):61-66.
[19]Majumdar A,Ward R K.Classification via group sparsity promoting regularization[C]//IEEE International Conference on Acoustics,Speech and Signal Processing.Taipei,Taiwan,China:IEEE,2009:861-864.
[20]Bach F,Jenatton R,Mairal J,et al.Structured sparsity through convex optimization[J].Statistical Science,2012,27(4):450-468.
[21]Bengio S,Pereira F,Singer Y,et al.Group sparse coding[C]//Advances in Neural Information Processing Systems.Whistler,Canada:MIT Press,2009:82-89.
[22]Chi Y T,Ali M,Rajwade A,et al.Block and group regularized sparse modeling for dictionary learning[C]//2013 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Oregon,Portland:IEEE,2013:377-382.
[23]Lei Hao,Mei Kuizhi,Zheng Nanning,et al.Learning group-based dictionaries for discriminative image representation[J].Pattern Recognition,2014,47(2):899-913.
[24]Liu Baodi,Wang Yuxiong,Shen Bin,et al.Blockwise coordinate descent schemes for efficient and effective dictionary learning[J].Neurocomputing,2016,178:25-35.
[25]Nie Feiping,Huang Heng,Cai Xiao,et al.Efficient and robust feature selection via joint 2,1-norms minimization[C]//Advances in Neural Information Processing Systems.Vancouver,Canada:MIT Press,2010:1813-1821.
[26]Peng Hanyang,Fan Yong.Direct l_(2,p)-norm learning for feature selection[J].arXiv preprint arXiv:1504.00430,2015.
[27]Wang Liping,Chen Songcan.Joint representation classification for collective face recognition[J].arXiv preprint arXiv:1505.04617,2015.
[28]Lee H,Battle A,Raina R,et al.Efficient sparse coding algorithms[C]//Advances in Neural Information Processing Systems.Vancouver,Canada:MIT Press,2006:801-808.
[29]Cai Sijia,Zuo Wangmeng,Zhang Lei,et al,Support vector guided dictionary learning[C]//European Conference on Computer Vision.Zurich:Springer,2014:624-639.

相似文献/References:

[1]刘 梓,宋晓宁,於东军,等.基于多成分字典和稀疏表示的超分辨率重建算法[J].南京理工大学学报(自然科学版),2014,38(01):1.
 Liu Zi,Song Xiaoning,Yu Dongjun,et al.Super-resolution reconstruction algorithm based on multi-component dictionary and sparse representation[J].Journal of Nanjing University of Science and Technology,2014,38(01):1.
[2]杨赛,赵春霞.基于空间概率乘积核函数的图像分类算法[J].南京理工大学学报(自然科学版),2014,38(03):325.
 Yang Sai,Zhao Chunxia.Image classification algorithm based on spatial probability product kernel[J].Journal of Nanjing University of Science and Technology,2014,38(01):325.
[3]封 磊,孙怀江.结合局部稀疏性和非局部相似性的盲压缩感知方法[J].南京理工大学学报(自然科学版),2017,41(04):399.[doi:10.14177/j.cnki.32-1397n.2017.41.04.001]
 Feng Lei,Sun Huaijiang.Blind compressive sensing method via local sparsity and nonlocal similarity[J].Journal of Nanjing University of Science and Technology,2017,41(01):399.[doi:10.14177/j.cnki.32-1397n.2017.41.04.001]

备注/Memo

备注/Memo:
收稿日期:2016-12-13 修回日期:2016-12-30
基金项目:国家自然科学基金(61373063)
作者简介:时中荣(1979-),女,博士生,主要研究方向:模式识别、稀疏分类,E-mail:shizrong@163.com; 通讯作者:刘传才(1963-),男,博士,教授,博士生导师,主要研究方向:模式识别、图像处理,E-mail:chuancailiu@njust.edu.cn。
引文格式:时中荣,王胜,刘传才.基于L2,p矩阵范数稀疏表示的图像分类方法[J].南京理工大学学报,2017,41(1):80-89.
投稿网址:http://zrxuebao.njust.edu.cn
更新日期/Last Update: 2017-02-28