Dual low-rank structure embedding for robust visual information processing

被引:1
作者
Zhou, Jianhang [1 ,2 ,3 ]
Zhang, Hengmin [4 ]
Li, Shuyi [5 ]
Zhang, Bob [6 ]
Fang, Leyuan [7 ]
Zhang, David [2 ,3 ]
机构
[1] Osaka Univ, Inst Sci & Ind Res, Dept Intelligent Media, Osaka 5670047, Japan
[2] Chinese Univ Hong Kong, Sch Data Sci, Shenzhen CUHK Shenzhen, Shenzhen 410082, Peoples R China
[3] Shenzhen Inst Artificial Intelligence & Robot Soc, Shenzhen 410082, Peoples R China
[4] Nanyang Technol Univ NTU, Sch Elect & Elect Engn, Singapore, Singapore
[5] Beijing Univ Technol, Fac Informat Technol, Beijing 100124, Peoples R China
[6] Univ Macau, Dept Comp & Informat Sci, Pattern Anal & Machine Intelligence Res Grp, Ave Univ, Taipa 999078, Macao, Peoples R China
[7] Hunan Univ, Coll Elect & Informat Engn, Changsha 410082, Peoples R China
基金
中国国家自然科学基金;
关键词
Low rank; Structure embedding; Bayesian inference; Visual information processing; SPARSE; RECOGNITION;
D O I
10.1016/j.knosys.2024.111821
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The low -rank (LR) property is widely applied to capture the global as well as intrinsic structure of the given data in different visual information processing tasks. Actually, there are three key information to determine the performance and generalization low -rank property based methods: (1) visual intrinsic structural information, (2) visual representation structural information, (3) visual robust information. To achieve these jointly in a unified framework, in this paper, we propose D ual L ow -rank S tructure E mbedding (DLSE) that embeds structural and robust information. We additionally proposed the J oint M atrix -based L inear R epresentation (JMLR) and theoretically proved it can realize DLSE. The proposed method was validated on 6 datasets (from 1,440 samples to 70,000 samples in size) and showed a promising performance in visual recognition (14.32% improvement compared with the deep features). In addition, we performed multiple analysis on robustness, representation, and its parameters to show the effectiveness of DLSE from different aspects.
引用
收藏
页数:12
相关论文
共 40 条
[1]   Efficient classification with sparsity augmented collaborative representation [J].
Akhtar, Naveed ;
Shafait, Faisal ;
Mian, Ajmal .
PATTERN RECOGNITION, 2017, 65 :136-145
[2]  
Azam B., 2022, Inform. Sci
[3]  
Balakrishnan N., 2007, Advances in Distribution Theory, Order Statistics, and Inference
[4]   Distributed optimization and statistical learning via the alternating direction method of multipliers [J].
Boyd S. ;
Parikh N. ;
Chu E. ;
Peleato B. ;
Eckstein J. .
Foundations and Trends in Machine Learning, 2010, 3 (01) :1-122
[5]   A Probabilistic Collaborative Representation based Approach for Pattern Classification [J].
Cai, Sijia ;
Zhang, Lei ;
Zuo, Wangmeng ;
Feng, Xiangchu .
2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, :2950-2959
[6]   Unequal adaptive visual recognition by learning from multi-modal data [J].
Cai, Ziyun ;
Zhang, Tengfei ;
Jing, Xiao-Yuan ;
Shao, Ling .
INFORMATION SCIENCES, 2022, 600 :1-21
[7]   Robust Principal Component Analysis? [J].
Candes, Emmanuel J. ;
Li, Xiaodong ;
Ma, Yi ;
Wright, John .
JOURNAL OF THE ACM, 2011, 58 (03)
[8]   Matrix Variate Distribution-Induced Sparse Representation for Robust Image Classification [J].
Chen, Jinhui ;
Yang, Jian ;
Luo, Lei ;
Qian, Jianjun ;
Xu, Wei .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2015, 26 (10) :2291-2300
[9]  
Clanuwat T, 2018, Arxiv, DOI arXiv:1812.01718
[10]  
Fang K.-T., 1992, Journal of the American Statistical Association, V87, P253, DOI [10.2307/2290483, DOI 10.2307/2290483]