计算机科学 ›› 2021, Vol. 48 ›› Issue (6): 86-95.doi: 10.11896/jsjkx.200800180

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于二阶近邻的核子空间聚类

王中元, 刘惊雷   

  1. 烟台大学计算机与控制工程学院 山东 烟台264005
  • 收稿日期:2020-08-27 修回日期:2020-09-18 出版日期:2021-06-15 发布日期:2021-06-03
  • 通讯作者: 刘惊雷(jinglei_liu@sina.com)
  • 基金资助:
    国家自然科学基金(61572419,61773331,61703360,61801414)

Kernel Subspace Clustering Based on Second-order Neighbors

WANG Zhong-yuan, LIU Jing-lei   

  1. School of Computer and Control Engineering,Yantai University,Yantai,Shandong 264005,China
  • Received:2020-08-27 Revised:2020-09-18 Online:2021-06-15 Published:2021-06-03
  • About author:WANG Zhong-yuan,born in 1996,postgraduate.His main research interests include kernel approximation of low rank block matrix and so on.(79186799@qq.com)
    LIU Jing-lei,born in 1970,Ph.D,professor,master supervisor.His main research interests include artificial intelligent and theoretical computer science.
  • Supported by:
    National Natural Science Foundation of China (61572419,61773331,61703360,61801414).

摘要: 高维数据集的处理是计算机视觉领域的核心,子空间聚类是实现高维数据聚类使用最广泛的方法之一。传统的子空间聚类假定数据来自不同的线性子空间,且不同子空间的区域不重叠。然而,现实中的数据往往不满足这两个约束条件,使得子空间聚类的效果受到影响。为了解决这两个问题,引入核化子空间来解决子空间数据的非线性问题,引入子空间系数矩阵的二阶近邻来处理重叠的子空间问题。随后,设计了基于二阶近邻的核化子空间三步聚类算法,首先求取核化子空间数据的自相似系数,然后消除子空间的重叠区域,最后对系数矩阵进行谱聚类。将所设计的子空间聚类算法首先在人工数据集上进行了测试,随后在人脸、场景字符和生物医学3类数据集中共12个真实数据集上进行了实验。实验结果表明,所提算法相比最新的几种算法具有一定的优势。

关键词: 二阶近邻, 核方法, 交替方向乘子法, 图像识别, 子空间聚类

Abstract: The processing of high-dimensional data sets is the focus of computer vision.Subspace clustering is one of the most widely used methods to achieve high-dimensional data clustering.The traditional subspace clustering assumes that the data comes from different linear subspaces,and different subspace regions do not overlap.However,real data often do not meet these two constraints,which affects the effect of subspace clustering.In order to deal with these two problems,this paper introduces a kernelized subspace to solve the nonlinear problem of subspace data,and introduces the second-order neighbors of the subspace coefficient matrix to deal with the overlapping subspace problem.Then a three-step clustering algorithm based on second-order neighbors of the kernelized subspace is designed.Firstly,the self-similarity coefficients of the kernelized subspace data are obtained.Secondly,the overlapping regions of the subspaces are eliminated.Finally,the coefficient matrix is spectrally clustered.In this paper,the designed subspace clustering algorithm is first tested on three artificial data sets,and then the experiment is performed on 12 real data sets,including face,scene characters and biomedical data sets.Experimental results show that the proposed algorithm has certain advantages over the latest algorithms.

Key words: Alternating direction multiplier method, Image identification, Kernel method, Second-order neighbors, Subspace clustering

中图分类号: 

  • TP18
[1]LIU Q S,LU H Q,MA S D.A review of subspace methods in face recognition Method[J].Acta Automatica Sinica,2003(6):900-911.
[2]DERKSEN H,YI M,WEI H.Segmentation of multivariatemixed data via lossy coding and compression[J].IEEE Tran-sactions on Pattern Analysis and Machine Intelligence,2007,29(9):1546-1562.
[3]ZHANG T,SZLAM A,LERMAN G.Median K-flats for hybrid linear modeling with many outliers [C]//IEEE 12th International Conference on Computer Vision Workshops.2009:234-241.
[4]HECKEL R,BOLCSKEI H.Robust Subspace Clustering viaThresholding [J].IEEE Transactions on Information Theory,2015,61(11):6320-6342.
[5]CAO W,ZHAO Y K,GAO S W.Multi-class support vector machine based on fuzzy kernel clustering[C]//Proceedings of 2009 China Process Systems Engineering Annual Conference and China Mes Annual Conference.2009:207-210.
[6]PENG X,YU Z,TANG H.Constructing the L2-Graph for Robust Subspace Learning and Subspace Clustering [J].IEEE Transactions on Cybernetics,2012,47(4):1053.
[7]DING Z C,GE H W,ZHOU J.Density peak clustering algorithm based on KL divergence[J].Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition),2019,31(3):367-374.
[8]ZHOU H,REN H P.Intuitionistic fuzzy similarity clustering algorithm based on chi-square distance[J].Journal of Chongqing University of Technology(Natural Science),2020,34(8):238-246.
[9]AGRAWAL R,GEHRKE J,GUNOPULOS D.Automatic subspace clustering of high dimensional data for data mining applications [J].ACM SIGMOD Record,1998,27(2):94-105.
[10]LU C,FENG J,YAN S.A Unified Alternating Direction Me-thod of Multipliers by Majorization Minimization[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,40(3):527-541.
[11]VIDAL R,FAVARO P.Low rank subspace clustering(LRSC) [J].Pattern Recognition Letters,2014,43(1):47-61.
[12]ELHAMIFAR E,VIDAL R.Sparse Subspace Clustering:Algorithm,Theory,and Applications [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(11):2765-2781.
[13]ZHANG X,CHEN X,SUN X.Schatten-q regularizer constrainedlow rank subspace clustering model [J].Neurocomputing,2016,182(C):36-47.
[14]YOU C,LI C G,ROBINSON D P.Oracle Based Active Set Algorithm for Scalable Elastic Net Subspace Clustering [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016.3928-3937.
[15]LU C Y,MIN H,ZHAO Z Q.Robust and Efficient SubspaceSegmentation via Least Squares Regression [C]//Proceedings ofthe 12th European Conference on Computer Vision-Volume Part VII.2012:347-360.
[16]CHONG Y,LI C G,ROBINSON D P.Oracle Based Active Set Algorithm for Scalable Elastic Net Subspace Clustering [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).2016:3928-3937.
[17]XU K J,RUAN J.Reweighted sparse subspace clustering [J].Computer Vision & Image Understanding,2015,138:25-37.
[18]ZHANG P T,CHEN X Y.Elastic nuclear subspace clustering[J].Pattern Recognition and Artificial Intelligence,2017,30(9):779-790.
[19]ZHANG Z,XU Y,SHAO L.Discriminative Block-DiagonalRepresentation Learning for Image Recognition[J].IEEE Transactions on Neural Networks & Learning Systems,2017,29(7):3111-3125.
[20]ABAVISANI M,PATEL V M.Multimodal sparse and low-rank subspace clustering [J].Information Fusion,2015,39:168-177.
[21]XIAO S,TAN M,XU D.Robust Kernel Low-Rank Representation [J].IEEE Transactions on Neural Networks & Learning Systems,2016,27(11):2268-2281.
[22]CHEN Y,JALALI A,SANGHAVI S.Clustering partially observed graphs via convex optimization [J].Journal of Machine Learning Research,2014,15(1):2213-2238.
[23]YANG J,LIANG J,WANG K.Subspace clustering via goodneighbors [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(6):1537-1544.
[24]CANDES E J,LI X,MA Y.Robust principal component analysis [J].Journal of the ACM,2011,58(3):11-48.
[25]LIU G,YAN S.Latent Low-Rank Representation for subspace segmentation and feature extraction [C]//InternationalConfe-rence on Computer Vision.2011:1615-1622.
[26]XIE X,GUO X,LIU G.Implicit Block Diagonal Low-Rank Representation [J].IEEE Transactions on Image Processing,2017,27(1):477-489.
[1] 王建明, 陈响育, 杨自忠, 史晨阳, 张宇航, 钱正坤.
不同数据增强方法对模型识别精度的影响
Influence of Different Data Augmentation Methods on Model Recognition Accuracy
计算机科学, 2022, 49(6A): 418-423. https://doi.org/10.11896/jsjkx.210700210
[2] 赵小虎, 叶圣, 李晓.
多算法融合的骨骼重建信息动作分类方法
Multi-algorithm Fusion Behavior Classification Method for Body Bone Information Reconstruction
计算机科学, 2022, 49(6): 269-275. https://doi.org/10.11896/jsjkx.210500070
[3] 徐建民, 孙朋, 吴树芳.
传播路径树核学习的微博谣言检测方法
Microblog Rumor Detection Method Based on Propagation Path Tree Kernel Learning
计算机科学, 2022, 49(6): 342-349. https://doi.org/10.11896/jsjkx.210400096
[4] 方仲礼, 王喆, 迟子秋.
面向多标签小样本学习的双流重构网络
Dual-stream Reconstruction Network for Multi-label and Few-shot Learning
计算机科学, 2022, 49(1): 212-218. https://doi.org/10.11896/jsjkx.201100143
[5] 熊朝阳, 王婷.
基于卷积神经网络的建筑构件图像识别
Image Recognition for Building Components Based on Convolutional Neural Network
计算机科学, 2021, 48(6A): 51-56. https://doi.org/10.11896/jsjkx.200500122
[6] 齐延荣, 周夏冰, 李斌, 周清雷.
基于FPGA的CNN图像识别加速与优化
FPGA-based CNN Image Recognition Acceleration and Optimization
计算机科学, 2021, 48(4): 205-212. https://doi.org/10.11896/jsjkx.200600089
[7] 石琳姗, 马创, 杨云, 靳敏.
基于SSC-BP神经网络的异常检测算法
Anomaly Detection Algorithm Based on SSC-BP Neural Network
计算机科学, 2021, 48(12): 357-363. https://doi.org/10.11896/jsjkx.201000086
[8] 顾兴健, 朱剑峰, 任守纲, 熊迎军, 徐焕良.
多尺度U网络实现番茄叶部病斑分割与识别
Multi-scale U Network Realizes Segmentation and Recognition of Tomato Leaf Disease
计算机科学, 2021, 48(11A): 360-366. https://doi.org/10.11896/jsjkx.201000166
[9] 刘洋, 金忠.
一种结合非局部和多区域注意力机制的细粒度图像识别方法
Fine-grained Image Recognition Method Combining with Non-local and Multi-region Attention Mechanism
计算机科学, 2021, 48(1): 197-203. https://doi.org/10.11896/jsjkx.191000135
[10] 高方远, 王秀美.
一种基于块对角表示和近邻约束的子空间聚类方法
Subspace Clustering Method Based on Block Diagonal Representation and Neighbor Constraint
计算机科学, 2020, 47(7): 66-70. https://doi.org/10.11896/jsjkx.190600155
[11] 邢毓华, 李明星.
基于投影的鲁棒低秩子空间聚类算法
Robust Low Rank Subspace Clustering Algorithm Based on Projection
计算机科学, 2020, 47(6): 92-97. https://doi.org/10.11896/jsjkx.190500074
[12] 刘淑君, 魏莱.
基于分块集成的图像聚类算法
Block Integration Based Image Clustering Algorithm
计算机科学, 2020, 47(6): 170-175. https://doi.org/10.11896/jsjkx.190400052
[13] 徐永士, 贲可荣, 王天雨, 刘斯杰.
DCGAN模型改进与SAR图像生成研究
Study on DCGAN Model Improvement and SAR Images Generation
计算机科学, 2020, 47(12): 93-99. https://doi.org/10.11896/jsjkx.200700109
[14] 朱颖雯, 杨君.
基于欧拉核的数据流聚类算法
Euler Kernel-based Data Stream Clustering Algorithm
计算机科学, 2019, 46(12): 74-82. https://doi.org/10.11896/jsjkx.190600158
[15] 杨秀璋, 夏换, 于小民.
一种基于水族濒危文字的图像增强及识别方法
Image Enhancement and Recognition Method Based on Shui-characters
计算机科学, 2019, 46(11A): 324-328.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!