Abstract:Several bandwise total variation (TV) regularized low-rank (LR)-based modelshave been proposed to remove mixed noise in hyperspectral images (HSIs).Conventionally, the rank of LR matrix is approximated using nuclear norm (NN).The NN is defined by adding all singular values together, which is essentiallya L_1-norm of the singular values. It results in non-negligible approximationerrors and thus the resulting matrix estimator can be significantly biased.Moreover, these bandwise TV-based methods exploit the spatial information in aseparate manner. To cope with these problems, we propose a spatial-spectral TV(SSTV) regularized non-convex local LR matrix approximation (NonLLRTV) methodto remove mixed noise in HSIs. From one aspect, local LR of HSIs is formulatedusing a non-convex L_γ-norm, which provides a closer approximation tothe matrix rank than the traditional NN. From another aspect, HSIs are assumedto be piecewisely smooth in the global spatial domain. The TV regularization iseffective in preserving the smoothness and removing Gaussian noise. These factsinspire the integration of the NonLLR with TV regularization. To address thelimitations of bandwise TV, we use the SSTV regularization to simultaneouslyconsider global spatial structure and spectral correlation of neighboringbands. Experiment results indicate that the use of local non-convex penalty andglobal SSTV can boost the preserving of spatial piecewise smoothness andoverall structural information.

Non-local extension of total variation regularization for image restoration

Nonlocal Gradient Sparsity Regularization for Image Restoration

Image Restoration Via Multi-Scale Non-Local Total Variation Regularization

Adaptive Non-local Total Variation for Image Restoration

An Efficient Image Reconstruction Framework Using Total Variation Regularization with Lp-Quasinorm and Group Gradient Sparsity

A New Difference of Anisotropic and Isotropic Total Variation Regularization Method for Image Restoration

Improved total variation based image compressive sensing recovery by nonlocal regularization

Image restoration using total variation with overlapping group sparsity

Video Restoration Based on a Novel Second Order Nonlocal Total Variation Model

Image Restoration by Combined Order Regularization with Optimal Spatial Adaptation

Image Denoising Via a Non-Local Patch Graph Total Variation

Hyperspectral Image Restoration Via Global Total Variation Regularized Local Nonconvex Low-Rank Matrix Approximation

A Tensor-Based Nonlocal Total Variation Model for Multi-Channel Image Recovery

Generalized Multi-Order Total Variation for Signal Restoration

Total Generalized Variation Restoration with Non-Quadratic Fidelity

High-order TVL1-based images restoration and spatially adapted regularization parameter selection.

Image restoration using overlapping group sparsity on hyper-Laplacian prior of image gradient

Compressive total variation for image reconstruction and restoration

Convex and non-convex adaptive TV regularizations for color image restoration

Total Variation Denoising With Non-Convex Regularizers.

Spatially adapted total variation model to remove multiplicative noise.