Fully-connected tensor network decomposition for robust tensor completion problem

DOI10.3934/IPI.2023030zbMATH Open1528.94007arXiv2110.08754OpenAlexW3206445549MaRDI QIDQ6087983FDOQ6087983

Authors: Yun-Yang Liu, Xi-Le Zhao, G. J. Song, Yu-Bang Zheng, Michael Ng, Ting-Zhu Huang 

Publication date: 13 December 2023

Published in: Inverse Problems and Imaging (Search for Journal in Brave)

Abstract: The robust tensor completion (RTC) problem, which aims to reconstruct a low-rank tensor from partially observed tensor contaminated by a sparse tensor, has received increasing attention. In this paper, by leveraging the superior expression of the fully-connected tensor network (FCTN) decomposition, we propose a $extbfFCTN$-based $extbfr$obust $extbfc$onvex optimization model (RC-FCTN) for the RTC problem. Then, we rigorously establish the exact recovery guarantee for the RC-FCTN. For solving the constrained optimization model RC-FCTN, we develop an alternating direction method of multipliers (ADMM)-based algorithm, which enjoys the global convergence guarantee. Moreover, we suggest a $extbfFCTN$-based $extbfr$obust $extbfn$on$extbfc$onvex optimization model (RNC-FCTN) for the RTC problem. A proximal alternating minimization (PAM)-based algorithm is developed to solve the proposed RNC-FCTN. Meanwhile, we theoretically derive the convergence of the PAM-based algorithm. Comprehensive numerical experiments in several applications, such as video completion and video background subtraction, demonstrate that proposed methods are superior to several state-of-the-art methods.

Full work available at URL: https://arxiv.org/abs/2110.08754

zbMATH Keywords

fully-connected tensor network decompositionrobust tensor completionexact recovery guarantee

Mathematics Subject Classification ID

Computational methods for sparse matrices (65F50) Nonconvex programming, global optimization (90C26) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08) Computing methodologies for image processing (68U10)

Cites Work

• Distributed optimization and statistical learning via the alternating direction method of multipliers
• Robust principal component analysis?
• The Power of Convex Relaxation: Near-Optimal Matrix Completion
• Tensor-train decomposition
• Hankel matrix rank minimization with applications to system identification and realization
• Factorization strategies for third-order tensors
• Proximal alternating linearized minimization for nonconvex and nonsmooth problems
• Parallel matrix factorization for low-rank tensor completion
• A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion
• Clarke Subgradients of Stratifiable Functions
• Third-order tensors as operators on matrices: a theoretical and computational framework with applications in imaging
• Most tensor problems are NP-hard
• Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods
• Tensor-Based Formulation and Nuclear Norm Regularization for Multienergy Computed Tomography
• A nonlocal low rank model for Poisson noise removal
• Low-rank tensor train for tensor robust principal component analysis
• Efficient Tensor Completion for Color Image and Video Recovery: Low-Rank Tensor Train
• Edge detection with mixed noise based on maximum a posteriori approach
• A new initialization method based on normed statistical spaces in deep networks
• Bilinear constraint based ADMM for mixed Poisson-Gaussian noise removal
• Tensor train rank minimization with nonlocal self-similarity for tensor completion
• Title not available (Why is that?)
• Nonlinear transform induced tensor nuclear norm for tensor completion
• Nonconvex optimization for robust tensor completion from grossly sparse observations
• Robust tensor completion using transformed tensor singular value decomposition
• Total variation based tensor decomposition for multi-dimensional data with time dimension.
• Guarantees of Riemannian optimization for low rank matrix completion
• Multi-dimensional image recovery via fully-connected tensor network decomposition under the learnable transforms
• Auto-weighted robust low-rank tensor completion via tensor-train
• Robust tensor train component analysis

Cited In (3)

• A random sampling algorithm for fully-connected tensor network decomposition with applications
• SVD-based algorithms for fully-connected tensor network decomposition
• Randomized tensor wheel decomposition

This page was built for publication: Fully-connected tensor network decomposition for robust tensor completion problem

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6087983)