Robust Tensor Principal Component Analysis by Lp-Norm for Image Analysis

被引：0
|
作者：
Tang, Ganyi [1 ]
Lu, Guifu [2 ]
Wang, Zhongqun [3 ]
Xie, Yukai [2 ]
机构：
[1] Anhui Polytech Univ, AHPU, Sch Comp & Informat, Wuhu, Peoples R China
[2] Anhui Polytech Univ, AHPU, Sch Comp Sci & Informat, Wuhu, Peoples R China
[3] Anhui Polytech Univ, AHPU, Sch Management Engn, Wuhu, Peoples R China
来源：
2016 2ND IEEE INTERNATIONAL CONFERENCE ON COMPUTER AND COMMUNICATIONS (ICCC) | 2016年
基金：
中国国家自然科学基金;
关键词：
tensor; principal component analysis ( PCA); TPCA; Lp-norm; outlies; MAXIMIZATION; L1-NORM;
D O I：
暂无
中图分类号：
TP3 [计算技术、计算机技术];
学科分类号：
0812 ;
摘要：
Tensor principal component analysis (TPCA), which can make full use of the spatial relationship of images/videos, is a generalization of the classical principal component analysis (PCA). However, the existing TPCA method is based on the Frobenius-norm, which makes it sensitive to outliers. In order to overcome the drawback of TPCA, in this paper, we proposed a novel Lp-norm based TPCA (TPCA-Lp), which is robust to outliers. We also designed an iterative algorithm to solve the optimization of TPCA-Lp, in which all projection matrices are optimized by turns. Experimental results upon several face databases demonstrate the advantages of the proposed approach.
引用
收藏
页码：568 / 573
页数：6
相关论文
共 50 条
  • [31] Tensor Robust Principal Component Analysis with Hankel Structure
    Xu, Chao
    Tan, Hao
    Feng, Qingrong
    Zhang, Yue
    Wang, Jianjun
    PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2024, PT 1, 2025, 15031 : 125 - 138
  • [32] A Sequential Lp-norm Filter for Robust Estimation
    Yang, Yang
    PROCEEDINGS OF 2020 23RD INTERNATIONAL CONFERENCE ON INFORMATION FUSION (FUSION 2020), 2020, : 1228 - 1232
  • [33] Graph-regularized tensor robust principal component analysis for hyperspectral image denoising
    Nie, Yongming
    Chen, Linsen
    Zhu, Hao
    Du, Sidan
    Yue, Tao
    Cao, Xun
    APPLIED OPTICS, 2017, 56 (22) : 6094 - 6102
  • [34] Generalization of linear discriminant analysis using Lp-norm
    Oh, Jae Hyun
    Kwak, Nojun
    PATTERN RECOGNITION LETTERS, 2013, 34 (06) : 679 - 685
  • [35] Incremental tensor principal component analysis for image recognition
    1600, Trans Tech Publications Ltd, Kreuzstrasse 10, Zurich-Durnten, CH-8635, Switzerland (710):
  • [36] Robust geometric lp-norm feature pooling for image classification and action recognition
    Li, Teng
    Meng, Zhijun
    Ni, Bingbing
    Shen, Jianbing
    Wang, Meng
    IMAGE AND VISION COMPUTING, 2016, 55 : 64 - 76
  • [37] Lp-Norm IDF for Scalable Image Retrieval
    Zheng, Liang
    Wang, Shengjin
    Tian, Qi
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2014, 23 (08) : 3604 - 3617
  • [38] A fast correction approach to tensor robust principal component analysis
    Zhang, Zhechen
    Liu, Sanyang
    Lin, Zhiping
    Xue, Jize
    Liu, Lixia
    APPLIED MATHEMATICAL MODELLING, 2024, 128 : 195 - 219
  • [39] Convex-Concave Tensor Robust Principal Component Analysis
    Liu, Youfa
    Du, Bo
    Chen, Yongyong
    Zhang, Lefei
    Gong, Mingming
    Tao, Dacheng
    INTERNATIONAL JOURNAL OF COMPUTER VISION, 2024, 132 (05) : 1721 - 1747
  • [40] Approximate Bayesian Algorithm for Tensor Robust Principal Component Analysis
    Srakar, Andrej
    NEW FRONTIERS IN BAYESIAN STATISTICS, BAYSM 2021, 2022, 405 : 1 - 9
12345