计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (10): 134-143.doi: 10.11896/jsjkx.240800076

• 计算机图形学&多媒体 • 上一篇    下一篇

基于稀疏协同相关熵的鲁棒主成分分析

陈平1, 刘珂菡2, 梁正友1, 胡奇兴2, 张远鹏3,4   

  1. 1 广西大学计算机、电子与信息学院 南宁 530000
    2 成都信息工程大学 成都 610225
    3 南通大学医学信息学系 江苏 南通 226000
    4 香港理工大学健康科技与资讯学系 香港 999077
  • 收稿日期:2024-08-14 修回日期:2024-12-05 出版日期:2025-10-15 发布日期:2025-10-14
  • 通讯作者: 梁正友(zhyliang@gxu.edu.cn)
  • 作者简介:(ppyuesheng@hotmail.com)
  • 基金资助:
    国家自然科学基金(12101090);四川省科技厅资助项目(2023NSFSC1425,2023NSFSC0071,2023NSFSC1362);2023成都信息工程大学科技创新能力提升计划创新团队重点项目(KYTD202322)

Sparsity Cooperated Correntropy Based Robust Principal Component Analysis

CHEN Ping1, LIU Kehan2, LIANG Zhengyou1, HU Qixing2, ZHANG Yuanpeng3,4   

  1. 1 School of Computer,Electronics and Information,Guangxi University,Nanning 530000,China
    2 Chengdu University of Information Technology,Chengdu 610225,China
    3 Department of Medical Informatics,Nantong University,Nantong,Jiangsu 226000,China
    4 Department of Health and Technology,Hong Kong Polytechnic University,Hong Kong 999077,China
  • Received:2024-08-14 Revised:2024-12-05 Online:2025-10-15 Published:2025-10-14
  • About author:CHEN Ping,born in 1986,postgra-duate.His main research interest is computer vision.
    LIANG Zhengyou,born in 1968,Ph.D,professor,is a member of CCF(No.16803M).His main research interests include computer vision and image processing.
  • Supported by:
    National Natural Science Foundation of China(12101090),Sichuan Provincial Science and Technology Department(2023NSFSC1425,2023NSFSC0071,2023NSFSC1362) and 2023 Chengdu University of Information Technology Science and Technology Innovation Capability Enhancement Plan Innovation Team Key Project(KYTD202322).

PDF (PC)

3

摘要： 主成分分析(Principal Component Analysis,PCA)广泛应用于许多领域,但其对非高斯噪声很敏感。研究者们已经提出了许多鲁棒主成分分析(Robust PCA,RPCA)模型来处理这个问题。然而,这些方法只能处理一种类型的噪声,如特征域中的脉冲噪声或样本域中的异常值。为此,提出了一种基于稀疏协同相关熵的RPCA模型(SCPCA),该模型对脉冲噪声和离群值同时具有鲁棒性。在此基础上,提出了一种基于Fenchel共轭和加速块坐标更新(Block Coordinate Update,BCU)策略的迭代算法。在聚类、背景重建和人脸建模方面进行了大量的实验来评估所提出的方法的鲁棒性。结果表明,在大多数情况下,所提出的方法优于目前先进的方法。

关键词: 鲁棒主成分分析, 相关熵, 背景重建, 人脸建模

Abstract: PCA is widely used in many applications but is sensitive to non-Gaussian noise.Many Robust PCA models have been proposed to handle this issue.However,these methods only can handle one type of noise,such as the impulse noise in the feature domain or the outliers in the sample domain.This paper proposes a novel RPCA model based on sparsity cooperated correntropy called SCPCA,which is robust against impulse noise and outlier simultaneously.Furthermore,an iterative algorithm is proposed to solve the proposed model based on the Fenchel conjugate and the accelerated BCU strategy.Extensive experiments on clustering,background reconstruction and face modelling have been conducted to evaluate the robustness of the proposed method.The results show that the proposed method outperforms the compared state of-the-art methods in most situations.

Key words: Robust principal component analysis,Correntropy,Background reconstruction,Face Modeling

中图分类号: 

  • TP181
[1]HOTELLING H.Analysis of a complex of statistical variables into principal components[J].Journal of Educational Psycho-logy,1933,24(6):417-441.
[2]RODARMEL C,SHAN J.Principal compo-nent analysis for hyperspectral image classification[J].Surveying and Land Information Science,2002,62(2):115-122.
[3]POON B,AMIN M A,YAN H.Pca based face recognition and testing criteria[C]//2009 International Conference on Machine Learning and Cybernetics.IEEE,2009:2945-2949.
[4]MA J,YUAN Y.Dimension reduction of image deep featureusing pca[J].Journal of Visual Communication and Image Representation,2019,63:102578.
[5]KWAK N.Principal component analysis based on l1-norm maximization[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(9):1672-1680.
[6]DING C,ZHOU D,HE X,et al.R1-PCA:rotational invariant l1-norm principal component analysis for robust sub space factorization[C]//Proceedings of the 23rd International Conference on Machine learning.2006:281-288.
[7]HE R,HU B G,ZHENG W S,et al.Robust principal componentanalysis based on maximum correntropy criterion[J].IEEE Transactions on Image Processing,2011,20(6):1485-1494.
[8]CANDES E J,LI X,MA Y,et al.Robust principal component analysis[J].Journal of the ACM,2011,58(3):1-37.
[9]WRIGHT J,GANESH A,RAO S,et al.Robust principal component analysis:Exact recovery of corrupted low-rank matrices via convex optimization[C]//Advances in Neural Information Processing Systems.2009.
[10]PENG Y,GANESH A,WRIGHT J,et al.Rasl:Robust align-ment by sparse and low-rank decom-position for linearly correlated images[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2233-2246.
[11]ZHOU T,TAO D.Godec:Randomized low-rank & sparse matrix decomposition in noisy case[C]//Proceedings of the 28th International Conference on Machine Learning.2011.
[12]GUO K,LIU L,XU X,et al.Godec+:Fast and robust low-rank matrix decom-position based on maximum correntropy[J].IEEE Transactions on Neural Networks and Learning Systems,2017,29(6):2323-2336.
[13]ZHAO Q,MENG D,XU Z,et al.Robust principal componentanalysis with complex noise[C]//International Conference on Machine Learning.PMLR,2014:55-63.
[14]LIU W,POKHAREL P P,PRINCIPE J C.Correntropy:Properties and applications in non-gaussian signal processing[J].IEEE Transactions on Signal Processing,2007,55(11):5286-5298.
[15]CHEN B,PRINCIPE J C.Maximum corr-entropy estimation is a smoothed map estimation[J].IEEE Signal Processing Letters,2012,19(8):491-494.
[16]DU L,LI X,SHEN Y D.Robust non negative matrix factorization via half-quadratic minimization[C]//2012 IEEE 12th International Conference on Data Mining.IEEE,2012:201-210.
[17]PENG S,CHEN B,SUN L,et al.Constrained maximum cor-rentropy adaptive filtering[J].Signal Processing,2017,140:116-126.
[18]ZHOU N,CHENG H,QIN J,et al.Robust high order manifold constrained sparse principal component analysis for image representation[J].IEEE Transactions on Circuits and Systems for Video Technology,2018,29(7):1946-1961.
[19]GEMAN D,REYNOIDS G.Constrained restoration and the recovery of discontinuities[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1992,14(3):367-383.
[20]XU Y,YIN W.A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion[J].SIAM Journal on Imaging Sciences,2013,6(3):1758-1789.
[21]SHI H J M,TU S,XU Y,et al.A primer on coordinate descent algorithms[J].arXiv:1610.00040,2016.
[22]XU Y.Alternating proximal gradient method for sparse nonnegative tucker decomposition[J].Mathematical Programming Computation,2015,7:39-70.
[23]WANG Y,JODOIN P M,PORIKLI F,et al.Cdnet 2014:An expanded change detection benchmark dataset[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops.2014:387-394.
[24]TOYAMA K,KRUMM J,BRUMITT B,et al.Wallflower:Principles and practice of background maintenance[C]//Proceedings of the 7th IEEE International Conference on Computer Vision.IEEE,1999:255-261.
[25]GEORGHIADES A S,BELHUMEUR P N, KRIEGMAN J D.From few to many:Illumination cone models for face recognition under variable lighting and pose[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2001,23(6):643-660.
[26]COLE R,FANTY M.ISOLET.UCI Machine Learning Repository[DB/OL].https://doi.org/-10.24432/C51G69.
[27]SILVERMAN B W.Density estimation for statistics and dataanalysis[M].CRC Press,1986.
[28]CHENG L,GONG M,SCHUURMANS D,et al.Real-time discriminative background subtraction[J].IEEE Transactions on Image Processing,2010,20(5):1401-1414.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发