Article
Geochemistry & Geophysics
Taner Ince, Nicolas Dobigeon
Summary: A weighted residual nonnegative matrix factorization method with spatial regularization is proposed to unmix hyperspectral data, aiming to improve the robustness of NMF against noise. Experimental results validate the effectiveness of the proposed method in providing spatial information for abundance matrix.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Ziyang Guo, Anyou Min, Bing Yang, Junhong Chen, Hong Li, Junbin Gao
Summary: This article proposes a new sparse oblique-manifold (OB) NMF method inspired by matrix manifold theory, treating the abundance matrix as located on the oblique manifold to eliminate constraints and incorporate intrinsic Riemannian geometry. By using the Riemannian conjugated gradient (RCG) algorithm and multiplicative iterative rule, the proposed method not only improves solution accuracy but also achieves a faster convergence rate. Experimental results demonstrate the effectiveness and efficiency of the proposed method compared to state-of-the-art NMF methods in HU.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Jin-Ju Wang, Ding-Cheng Wang, Ting-Zhu Huang, Jie Huang, Xi-Le Zhao, Liang-Jian Deng
Summary: In this paper, a new NTF-based model called EIC-NTF is proposed for hyperspectral unmixing to mitigate the impact of high correlation among endmembers and abundances. The model introduces endmember independence constraint for endmember estimation and exploits the low-rankness in abundance maps for abundance estimation. Experimental results show that the proposed algorithm is effective for hyperspectral unmixing.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Geochemistry & Geophysics
Fengchao Xiong, Jun Zhou, Shuyin Tao, Jianfeng Lu, Yuntao Qian
Summary: This article presents an end-to-end deep neural network named SNMF-Net for hyperspectral unmixing, which combines model-based and learning-based methods. SNMF-Net shows high physical interpretability and performance advantages over state-of-the-art methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Qin Jiang, Yifei Dong, Jiangtao Peng, Mei Yan, Yi Sun
Summary: The paper introduces a robust MLE-based NMF model for hyperspectral unmixing, which shows superior performance compared to existing NMF methods in experiments using simulated and real hyperspectral data sets.
Article
Computer Science, Artificial Intelligence
Nicolas Nadisic, Jeremy E. Cohen, Arnaud Vandaele, Nicolas Gillis
Summary: This paper introduces a new form of sparse MNNLS problem and a two-step algorithm to solve it. By dividing the problem into subproblems and selecting Pareto front solutions, a matrix that satisfies the sparsity constraint is constructed. Experimental results show that this method is more accurate than existing heuristic algorithms.
Article
Operations Research & Management Science
Takehiro Sano, Tsuyoshi Migita, Norikazu Takahashi
Summary: In this paper, a novel update rule for the HALS algorithm is proposed, which is well-defined and guarantees global convergence. The proposed rule allows sparse factor matrices by allowing variables to take the value of zero. Two stopping conditions are also presented to ensure the finite termination of the algorithm.
JOURNAL OF GLOBAL OPTIMIZATION
(2022)
Article
Geochemistry & Geophysics
Xuelong Li, Xinxin Zhang, Yuan Yuan, Yongsheng Dong
Summary: Hyperspectral unmixing is an important research topic for spectral data analysis. Existing methods often treat similarity learning and unmixing as two separate steps, resulting in suboptimal similarity matrices. To address this issue, we propose an adaptive relationship preserving-based sparse NMF method that improves the performance and generalization ability of unmixing.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Risheng Huang, Huiyun Jiao, Xiaorun Li, Shuhan Chen, Chaoqun Xia
Summary: In this study, a robust deep nonnegative matrix factorization (l(2,1)-RDNMF) based on the l(2,1) norm is proposed for hyperspectral unmixing. The l(2,1)-RDNMF incorporates the l(2,1) norm into the deep structure to achieve robustness. The efficiency and effectiveness of the proposed method are verified through experiments using synthetic and genuine data.
Article
Mathematics, Applied
Yingying Xu, Lei Du, Haifeng Song, Songsong Dai
Summary: The significance of hyperspectral unmixing lies in separating mixed spectral signatures into endmembers and calculating their fractional abundances. However, blind unmixing based on non-negative matrix factorization (NMF) often yields inaccurate and unstable results, while semi-blind unmixing based on sparse regression requires an overcomplete spectral library. To overcome these limitations, a cooperative model called CoNMFSL is proposed, which combines NMF and spectral library to utilize partial prior information from the incomplete spectral library. Extensive simulations and real-data experiments demonstrate that CoNMFSL achieves more accurate and stable results compared to blind unmixing, and relaxes the application conditions of semi-blind unmixing.
JOURNAL OF NONLINEAR AND CONVEX ANALYSIS
(2022)
Article
Computer Science, Artificial Intelligence
Yuan Yuan, Zihan Zhang, Ganchao Liu
Summary: Hyperspectral remote sensing is an important method for earth observation, but the low spatial resolution of images makes it hard to distinguish ground objects. The HP-MLNMF framework proposed in this paper shows promising results in improving unmixing speed and efficiency in spectral signature estimation.
Article
Geochemistry & Geophysics
Jiangtao Peng, Weiwei Sun, Fan Jiang, Hong Chen, Yicong Zhou, Qian Du
Summary: This letter proposes a general loss-based NMF (GLNMF) model for hyperspectral unmixing. By introducing a robust loss function, the model improves its robustness and achieves higher accuracy compared to existing NMF methods, as demonstrated through experimental results on simulated and real data sets.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Min Zhao, Tiande Gao, Jie Chen, Wei Chen
Summary: This study introduces an NMF-based unmixing framework that combines learned regularizers and handcrafted regularizers, demonstrating the integration of handcrafted regularizers using l2,1-norm as an example. The proposed framework is flexible and extendable, with effectiveness confirmed through both synthetic and real airborne data.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Environmental Sciences
Xinxi Feng, Le Han, Le Dong
Summary: This paper proposes a non-negative tensor factorization framework based on weighted group sparsity constraint for the unmixing of hyperspectral images. By using the weighted constraint of the L2,1 norm, the method can explore the similar characteristics in the spectral dimension and maintain data smoothness in the spatial dimension.
Article
Geochemistry & Geophysics
Jiangtao Peng, Yicong Zhou, Weiwei Sun, Qian Du, Lekang Xia
Summary: This article introduces three self-paced learning based NMF unmixing models that use weighted least-squares losses instead of the least-squares loss in the standard NMF model. By adopting a self-paced learning strategy to adaptively learn weights, the SpNMF models demonstrate better accuracy and robustness in handling different types of noise in hyperspectral data. Gradually enlarging the model set and selecting atoms from easy to complex allows SpNMF to exclude noisy or outlying atoms, resulting in a more robust unmixing model.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)