8 Hits in 5.1 sec

Two-Stage Nonnegative Sparse Representation for Large-Scale Face Recognition

Ran He, Wei-Shi Zheng, Bao-Gang Hu, Xiang-Wei Kong
2013 IEEE Transactions on Neural Networks and Learning Systems  
Based on the divide and conquer strategy, TSR decomposes the procedure of robust face recognition into outlier detection stage and recognition stage.  ...  This paper proposes a novel nonnegative sparse representation approach, called two-stage sparse representation (TSR), for robust face recognition on a large-scale database.  ...  [32] adopted robust stochastic approximation for online nonnegative factorization.  ... 
doi:10.1109/tnnls.2012.2226471 pmid:24808205 fatcat:l4zt22t6nbfwdbsb4npa6j7ln4

Robust Nonnegative Matrix Factorization via Half-Quadratic Minimization

Liang Du, Xuan Li, Yi-Dong Shen
2012 2012 IEEE 12th International Conference on Data Mining  
Nonnegative matrix factorization (NMF) is a popular technique for learning parts-based representation and data clustering.  ...  In this paper, we propose a robust NMF method based on the correntropy induced metric, which is much more insensitive to outliers.  ...  ROBUST NONNEGATIVE MATRIX FACTORIZATION In this section, we will derive three robust NMFs, which use the Correntropy Induced Metric or the Huber Mestimator to measure the quality of matrix approximation  ... 
doi:10.1109/icdm.2012.39 dblp:conf/icdm/DuLS12 fatcat:ojlkd6vy4zdmphihi7whbslf2u

Robust and Low-Rank Representation for Fast Face Identification With Occlusions

Michael Iliadis, Haohong Wang, Rafael Molina, Aggelos K. Katsaggelos
2017 IEEE Transactions on Image Processing  
Our approach utilizes a robust representation based on two characteristics in order to model contiguous errors (e.g., block occlusion) effectively.  ...  In this paper we propose an iterative method to address the face identification problem with block occlusions.  ...  In correntropy-based sparse representation (CESR) [12] and structured sparse error coding (SSEC) [15] , ϑ(a) was chosen to be the indicator function of the nonnegative orthant R n + , such that a nonnegative  ... 
doi:10.1109/tip.2017.2675206 pmid:28252401 fatcat:d4hycbprrvf7rl4th5hmw6rkz4

Multi-model robust error correction for face recognition

Michael Iliadis, Leonidas Spinoulas, Albert S. Berahas, Haohong Wang, Aggelos K. Katsaggelos
2016 2016 IEEE International Conference on Image Processing (ICIP)  
Finally, the multi-model residual representation offers useful insights into understanding how different noise types affect face recognition rates.  ...  In this work we present a general framework for robust error estimation in face recognition.  ...  In correntropy-based sparse representation (CESR) [4] , ϑ(a) was chosen to be the indicator function of the nonnegative orthant R n + , such that a nonnegative a ≥ 0 regularization term was enforced.  ... 
doi:10.1109/icip.2016.7532956 dblp:conf/icip/IliadisSBWK16 fatcat:hg5adzzfrvhfzab7fxtlvqhrzi

Truncated Cauchy Non-negative Matrix Factorization for Robust Subspace Learning

Naiyang Guan, Tongliang Liu, Yangmuzi Zhang, Dacheng Tao, Larry Steven Davis
2017 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Non-negative matrix factorization (NMF) minimizes the Euclidean distance between the data matrix and its low rank approximation, and it fails when applied to corrupted data because the loss function is  ...  In this paper, we propose a Truncated Cauchy nonnegative matrix factorization (Truncated CauchyNMF) model to learn a subspace on a dataset contaminated by large magnitude noise or corruption.  ...  Correntropy Induced Metric Based NMF The most closely-related work is the half-quadratic algorithm for optimizing robust NMF, which includes the Correntropy-Induced Metric (CIM)-based NMF (CIM-NMF) and  ... 
doi:10.1109/tpami.2017.2777841 pmid:29990056 fatcat:v7ornfymyfcvbdvaanrjpk3c6e

2021 Index IEEE Transactions on Image Processing Vol. 30

2021 IEEE Transactions on Image Processing  
Wang, C., +, TIP 2021 7980-7994 Blind Decomposition of Multispectral Document Images Using Orthogonal Nonnegative Matrix Factorization.  ...  Huang, Y., +, TIP 2021 2325-2339 Document image processing Blind Decomposition of Multispectral Document Images Using Orthogonal Nonnegative Matrix Factorization.  ... 
doi:10.1109/tip.2022.3142569 fatcat:z26yhwuecbgrnb2czhwjlf73qu

Robust Learning Based on The Information Theoretic Learning

Miaohua Zhang, University, My, Yongsheng Gao
Since the correntropy and generalized correntropy are both based [...]  ...  However, the MCC-based loss function uses the second order measurement to constraint the representation error, which is not always the best choice.  ...  In regression analysis based face recognition methods, we use training images to represent a test image. Ideally, the error image is a zero matrix.  ... 
doi:10.25904/1912/3265 fatcat:lqifh6vpb5ch5dw3laibal32dy

ICWMC 2017 Committee ICWMC Steering Committee ICWMC 2017 Technical Program Committee

Carlos Westphall, Khalil El-Khatib, Hamid Menouar, Carlos Westphall, Afrand Agah, West Chester, Pradipta De, David Sanchez, Carl Debono, Augusto Morales, Cheng Bo, Huawei Technologies (+64 others)
We are convinced that the participants found the event useful and communications very open.  ...  The most widely used kernel in correntropy is the complex Gaussian kernel which is given by κ σ (ζ) = 1 √ 2πσ exp −|ζ| 2 2σ 2 (5) Comparing correntropy with MSE, we note that Correntropy is a local similarity  ...  As we can see, the type of car is another factor used to evaluate the speed.  ...