Hessian Schatten-Norm Regularization for Linear Inverse Problems

Abstract

We introduce a novel family of invariant, convex, and non-quadratic functionals that we employ to derive regularized solutions of ill-posed linear inverse imaging problems. The proposed regularizers involve the Schatten norms of the Hessian matrix, computed at every pixel of the image. They can be viewed as second-order extensions of the popular total-variation (TV) semi-norm since they satisfy the same invariance properties. Meanwhile, by taking advantage of second-order derivatives, they avoid the staircase effect, a common artifact of TV-based reconstructions, and perform well for a wide range of applications. To solve the corresponding optimization problems, we propose an algorithm that is based on a primal-dual formulation. A fundamental ingredient of this algorithm is the projection of matrices onto Schatten norm balls of arbitrary radius. This operation is performed efficiently based on a direct link we provide between vector projections onto ${\ell }_q$ norm balls and matrix projections onto Schatten norm balls. Finally, we demonstrate the effectiveness of the proposed methods through experimental results on several inverse imaging problems with real and simulated data.

15 pages double-column format. This manuscript will appear in IEEE Transactions on Image Processing

References in corpus (1)

• A Singular Value Thresholding Algorithm for Matrix Completion

Cited by in corpus (22)

• Tensor-based formulation and nuclear norm regularization for multi-energy computed tomography
• Iterative Joint Image Demosaicking and Denoising using a Residual Denoising Network
• Efficient Inversion of Multiple-Scattering Model for Optical Diffraction Tomography
• Robust Phase Unwrapping via Deep Image Prior for Quantitative Phase Imaging
• Higher-order total variation approaches and generalisations
• Pocket Guide to Solve Inverse Problems with GlobalBioIm
• On Plug-and-Play Regularization using Linear Denoisers
• SGD-Net: Efficient Model-Based Deep Learning with Theoretical Guarantees
• Isotropic inverse-problem approach for two-dimensional phase unwrapping
• Inverse design of gradient-index volume multimode converters
• Image Restoration by Combined Order Regularization with Optimal Spatial Adaptation
• Weight-Covariance Alignment for Adversarially Robust Neural Networks
• RHAPSODIE: Reconstruction of High-contrAst Polarized SOurces and Deconvolution for cIrcumstellar Environments
• Adaptive Direction-Guided Structure Tensor Total Variation
• Sensitivity-Aware Density Estimation in Multiple Dimensions
• Image Reconstruction from Undersampled Confocal Microscopy Data using Multiresolution Based Maximum Entropy Regularization
• Generalized Multi-Order Total Variation for Signal Restoration
• Tensor Based Second Order Variational Model for Image Reconstruction
• Structurally Adaptive Multi-Derivative Regularization for Image Recovery from Sparse Fourier Samples
• Diffraction Tomography with Helmholtz Equation: Efficient and Robust Multigrid-Based Solver
• Recursive vectorized computation of the vector $p$-norm
• Optimizing Graph Structure for Targeted Diffusion