Abstract:Tensor decompositions have been successfully applied to multidimensional data recovery. However, classical tensor decompositions are not suitable for emerging spatio-irregular multidimensional data (i.e., spatio-irregular tensor), whose spatial domain is non-rectangular, e.g., spatial transcriptomics data from bioinformatics and semantic units from computer vision. By using preprocessing (e.g., zero-padding or element-wise 0-1 weighting), the spatio-irregular tensor can be converted to a spatio-regular tensor and then classical tensor decompositions can be applied, but this strategy inevitably introduces bias information, leading to artifacts. How to design a tensor-based method suitable for emerging spatio-irregular tensors is an imperative challenge. To address this challenge, we propose a learnable transform-assisted tensor singular value decomposition (LTA-TSVD) for spatio-irregular tensor recovery, which allows us to leverage the intrinsic structure behind the spatio-irregular tensor. Specifically, we design a learnable transform to project the original spatio-irregular tensor into its latent spatio-regular tensor, and then the latent low-rank structure is captured by classical TSVD on the resulting regular tensor. Empowered by LTA-TSVD, we develop spatio-irregular low-rank tensor completion (SIR-LRTC) and spatio-irregular tensor robust principal component analysis (SIR-TRPCA) models for the spatio-irregular tensor imputation and denoising respectively, and we design corresponding solving algorithms with theoretical convergence. Extensive experiments including the spatial transcriptomics data imputation and hyperspectral image denoising show SIR-LRTC and SIR-TRPCA are superior performance to competing approaches and benefit downstream applications.

High-Order Tensor Recovery Coupling Multilayer Subspace Priori with Application in Video Restoration

Tensor Nonconvex Unified Prior For Tensor Recovery

Handling The Non-Smooth Challenge in Tensor SVD: A Multi-Objective Tensor Recovery Framework

Novel Factorization Strategies for Higher Order Tensors: Implications for Compression and Recovery of Multi-linear Data

Multiplex Transformed Tensor Decomposition for Multidimensional Image Recovery.

Multi-Dimensional Visual Data Completion via Low-Rank Tensor Representation Under Coupled Transform

Novel methods for multilinear data completion and de-noising based on tensor-SVD

Guaranteed Tensor Recovery Fused Low-rankness and Smoothness

An Effective Approach for Color Video Recovery Based on Balanced Tensor

Visual Data Completion Via Local Sensitive Low Rank Tensor Learning.

Low-Rank Tensor Learning with Discriminant Analysis for Action Classification and Image Recovery

Superpixel-based Robust Tensor Low-Rank Approximation for Multimedia Data Recovery

Low-Rank Tensor Completion with Spatio-Temporal Consistency

Parallel Matrix Factorization for Low-Rank Tensor Completion

Adaptive Dimension-Discriminative Low-Rank Tensor Recovery for Computational Hyperspectral Imaging.

Mixture Augmented Lagrange Multiplier Method for Tensor Recovery

Learnable Transform-Assisted Tensor Decomposition for Spatio-Irregular Multidimensional Data Recovery

Local-and-Nonlocal Spectral Prior Regularized Tensor Recovery for Cauchy Noise Removal

Multiscale Feature Tensor Train Rank Minimization for Multidimensional Image Recovery

Tensor Completion Using Bilayer Multimode Low-Rank Prior and Total Variation

Video SAR Imaging Based on Low-Rank Tensor Recovery