Robust canonical correlation analysis based on L1-norm minimization for feature learning and image recognition

Wang, Sheng; Du, Haishun; Zhang, Ge; Lu, Jianfeng; Yang, Jingyu

doi:10.1117/1.jei.29.2.023001

Cited by 3 publications

(5 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…21) can be accurately approximated by solving problem (Eq. 22), and the solution of l 1 -optimization problem is more convenient (Wang et al, 2020). It can be solved by the orthogonal matching pursuit (OMP) algorithm (Papadopoulos et al, 2019).…”

Section: Solution Flow Of Traditional Compressive Sensing Algorithmmentioning

confidence: 99%

Probabilistic power flow calculation using principal component analysis-based compressive sensing

et al. 2023

View full text Add to dashboard Cite

The increasing scale of the injection of renewable energy has brought about great uncertainty to the operation of power grid. In this situation, probabilistic power flow (PPF) calculation has been introduced to mitigate the low accuracy of traditional deterministic power flow calculation in describing the operation status and power flow distribution of power systems. Polynomial chaotic expansion (PCE) method has become popular in PPF analysis due to its high efficiency and accuracy, and sparse PCE has increased its capability of tackling the issue of dimension disaster. In this paper, we propose a principal component analysis-based compressive sensing (PCA-CS) algorithm solve the PPF problem. The l1-optimization of CS is used to tackle the dimension disaster of sparse PCE, and PCA is included to further increase the sparsity of expansion coefficient matrix. Theoretical and numerical simulation results show that the proposed method can effectively improve the efficiency of PPF calculation in the case of random inputs with higher dimensions.

show abstract

Section: Solution Flow Of Traditional Compressive Sensing Algorithmmentioning

confidence: 99%

Probabilistic power flow calculation using principal component analysis-based compressive sensing

et al. 2023

View full text Add to dashboard Cite

show abstract

“…When the parameters are changed, our method will obtain different results. We explore the influence of σ by varying it from 2 −5 to 2 5 and plot the clustering accuracy (ACC) of MvLRFD on the ORL dataset in Figure 5(a) with λ � 2 1 , β � 2 −4 . From equation ( 22), we know that σ has two important properties.…”

Section: Parameter Sensitivity Analysis and Convergencementioning

confidence: 99%

“…is pattern is explainable since a higher value of λ will force MvLRFD to focus more on fused information between views. for each iv ∈ [1, n v ] do (5) Computer the basis of i-th view using ( 16); (6) Computer latent representation of v-th view using ( 19); (7) Update the Dive using Dive � Dive + HY (i) (Y (i) ) T H; (8) end for (9) Update weight vector α using ( 22); (10) Update S using ( 29); (11) Compute the value of the objective function J t using ( 13); (12) if (|J t − J| < 1e − 3)‖(iter > maxIter) then (13) break; ( 14) else (15) J⟵J t ; (16) end if (17) iter⟵iter + 1; (18) end for (19) return the final representation using (8); ALGORITHM 1: e algorithm of MvLRFD. With σ � 2 2 and λ � 2 2 , our method will obtain a better performance for maintaining private information when 2 − 2 ≥ β ≥ 2 − 4 .…”

Section: Parameter Sensitivity Analysis and Convergencementioning

confidence: 99%

“…Recently, the multiview image clustering task has received increasing attention. Approximately, they can be roughly split into three classes: (1) CCA-based methods [3][4][5]; (2) matrix factorization-based methods [6,7]; (3) subspace clustering learning methods [8,9]. Canonical Correlation Analysis gets two mappings by making the projected data linearly correlated maximally.…”

Section: Introductionmentioning

confidence: 99%

“…From the corresponding reference, CCA-based multiview representation learning algorithms get nice performance in multiview clustering. However, by nature, CCA assumes that the paired data are the same as each other in the latent space [3,5], and CCA-based methods ignore the di erent fundamental attributes between views.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multiview Latent Representation Learning with Feature Diversity for Clustering

Wang

Chen

Peng³

2022

Mathematical Problems in Engineering

Self Cite

View full text Add to dashboard Cite

To analyze and manage data, multiple features are extracted for robust and accurate description. Taking account of the complementary nature of multiview data, it is urgent to find a low-dimensional compact representation by leveraging multiple views and generating a better performance for clustering. In this paper, we present multiview latent representation learning with feature diversity for clustering (MvLRFD). To get rid of noise and maintain the information of each view, matrix factorization is adopted to get the latent representation. For fusing the diverse latent representations, we argue that the final representation is constructed by concatenating the latent representation of each view. With the help of the Hilbert–Schmidt Independence Criterion (HSIC), the diversity of features in the final representation is maximized to exploit the complementary information. A new coregularization strategy is introduced to maximize the distance between the similarity between data of the final representation and the latent one. The partition entropy regularization is adopted to control the uniformity level of the value of the similarity matrix and the weight of each view. To verify the availability of MvLRFD compared to state-of-the-art algorithms, several experiments are conducted on four real datasets.

show abstract

Iris-Fingerprint multimodal biometric system based on optimal feature level fusion model

Kamlaskar¹,

Abhyankar

2021

ELECTRENG

View full text Add to dashboard Cite

<abstract><p>For reliable and accurate multimodal biometric based person verification, demands an effective discriminant feature representation and fusion of the extracted relevant information across multiple biometric modalities. In this paper, we propose feature level fusion by adopting the concept of canonical correlation analysis (CCA) to fuse Iris and Fingerprint feature sets of the same person. The uniqueness of this approach is that it extracts maximized correlated features from feature sets of both modalities as effective discriminant information within the features sets. CCA is, therefore, suitable to analyze the underlying relationship between two feature spaces and generates more powerful feature vectors by removing redundant information. We demonstrate that an efficient multimodal recognition can be achieved with a significant reduction in feature dimensions with less computational complexity and recognition time less than one second by exploiting CCA based joint feature fusion and optimization. To evaluate the performance of the proposed system, Left and Right Iris, and thumb Fingerprints from both hands of the SDUMLA-HMT multimodal dataset are considered in this experiment. We show that our proposed approach significantly outperforms in terms of equal error rate (EER) than unimodal system recognition performance. We also demonstrate that CCA based feature fusion excels than the match score level fusion. Further, an exploration of the correlation between Right Iris and Left Fingerprint images (EER of 0.1050%), and Left Iris and Right Fingerprint images (EER of 1.4286%) are also presented to consider the effect of feature dominance and laterality of the selected modalities for the robust multimodal biometric system.</p></abstract>

show abstract

Robust canonical correlation analysis based on L1-norm minimization for feature learning and image recognition

Cited by 3 publications

References 42 publications

Probabilistic power flow calculation using principal component analysis-based compressive sensing

Probabilistic power flow calculation using principal component analysis-based compressive sensing

Multiview Latent Representation Learning with Feature Diversity for Clustering

Iris-Fingerprint multimodal biometric system based on optimal feature level fusion model

Contact Info

Product

Resources

About