Smoothed separable nonnegative matrix factorization

DOI10.1016/J.LAA.2023.07.013MaRDI QIDQ6178776zbMATH OpenOpenAlexFDO

Authors Nicolas Nadisic, Nicolas Gillis, Christophe Kervazo

Publication date 5 September 2023

Published in Linear Algebra and its Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2110.05528

zbMATH Keywords

blind hyperspectral unmixinglatent simplexnonnegative matrix factorizationpure-pixel search algorithmsseparabilitysimplex-structured matrix factorization

Mathematics Subject Classification ID

Factorization of matrices (15A23) Numerical methods for low-rank matrix approximation; matrix compression (65F55) Computing methodologies for image processing (68U10)

Abstract: Given a set of data points belonging to the convex hull of a set of vertices, a key problem in linear algebra, signal processing, data analysis and machine learning is to estimate these vertices in the presence of noise. Many algorithms have been developed under the assumption that there is at least one nearby data point to each vertex; two of the most widely used ones are vertex component analysis (VCA) and the successive projection algorithm (SPA). This assumption is known as the pure-pixel assumption in blind hyperspectral unmixing, and as the separability assumption in nonnegative matrix factorization. More recently, Bhattacharyya and Kannan (ACM-SIAM Symposium on Discrete Algorithms, 2020) proposed an algorithm for learning a latent simplex (ALLS) that relies on the assumption that there is more than one nearby data point to each vertex. In that scenario, ALLS is probalistically more robust to noise than algorithms based on the separability assumption. In this paper, inspired by ALLS, we propose smoothed VCA (SVCA) and smoothed SPA (SSPA) that generalize VCA and SPA by assuming the presence of several nearby data points to each vertex. We illustrate the effectiveness of SVCA and SSPA over VCA, SPA and ALLS on synthetic data sets, on the unmixing of hyperspectral images, and on feature extraction on facial images data sets. In addition, our study highlights new theoretical results for VCA.

Cites work

• A Fast Hyperplane-Based Minimum-Volume Enclosing Simplex Algorithm for Blind Hyperspectral Unmixing
• Blind Separation of Quasi-Stationary Sources: Exploiting Convex Geometry in Covariance Domain
• Column subset selection is NP-complete
• Computing a nonnegative matrix factorization -- provably
• Finding a latent k–simplex in O* (k · nnz(data)) time via Subset Smoothing
• Learning the parts of objects by non-negative matrix factorization
• Maximum Volume Inscribed Ellipsoid: A New Simplex-Structured Matrix Factorization Framework via Facet Enumeration and Convex Optimization
• Nonnegative matrix factorization for spectral data analysis
• Principal component analysis.
• Robust near-separable nonnegative matrix factorization using linear optimization
• Robustness analysis of Hottopixx, a linear programming model for factoring nonnegative matrices
• Simplex-Structured Matrix Factorization: Sparsity-Based Identifiability and Provably Correct Algorithms
• Sparse nonnegative matrix underapproximation and its application to hyperspectral image analysis
• Successive nonnegative projection algorithm for robust nonnegative blind source separation

Cited in

(1)

• On the robustness of the successive projection algorithm

This page was built for publication: Smoothed separable nonnegative matrix factorization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6178776)