Implicit regularization effects of the Sobolev norms in image processing
From MaRDI portal
Sobolev (and similar kinds of) spaces of functions on metric spaces; analysis on metric spaces (46E36) Inverse problems in optimal control (49N45) Numerical optimization and variational techniques (65K10) Computing methodologies for image processing (68U10) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08)
Abstract: In this paper, we propose to use the general -based Sobolev norms, i.e., norms where , to measure the data discrepancy due to noise in image processing tasks that are formulated as optimization problems. As opposed to a popular trend of developing regularization methods, we emphasize that an implicit regularization effect can be achieved through the class of Sobolev norms as the data-fitting term. Specifically, we analyze that the implicit regularization comes from the weights that the norm imposes on different frequency contents of an underlying image. We further analyze the underlying noise assumption of using the Sobolev norm as the data-fitting term from a Bayesian perspective, build the connections with the Sobolev gradient-based methods and discuss the preconditioning effects on the convergence rate of the gradient descent algorithm, leading to a better understanding of functional spaces/metrics and the optimization process involved in image processing. Numerical results in full waveform inversion, image denoising and deblurring demonstrate the implicit regularization effects.
Recommendations
- Tikhonov regularization using Sobolev metrics
- Infimal convolution regularisation functionals of BV and \(\mathrm{L}^p\) spaces. I: The finite \(p\) case
- New possibilities in image diffusion and sharpening via high-order Sobolev gradient flows
- Linear image reconstruction by Sobolev norms on the bounded domain
- Analysis of the bounded variation and the G regularization for nonlinear inverse problems
Cites work
- A duality based approach to the minimizing total variation flow in the space \(H^{-s}\)
- A general framework for a class of first order primal-dual algorithms for convex optimization in imaging science
- A Nonlinear Primal-Dual Method for Total Variation-Based Image Restoration
- A Review of Image Denoising Algorithms, with a New One
- A Statistical Model for Positron Emission Tomography
- A Total Fractional-Order Variation Model for Image Restoration with Nonhomogeneous Boundary Conditions and Its Numerical Solution
- A variational approach to additive image decomposition into structure, harmonic, and oscillatory components
- A Variational Approach to Removing Multiplicative Noise
- A weighted difference of anisotropic and isotropic total variation model for image processing
- An algorithm for total variation minimization and applications
- An image sharpening operator combined with framelet for image deblurring
- An Iterative Regularization Method for Total Variation-Based Image Restoration
- Anderson acceleration based on the \(\mathcal{H}^{- s}\) Sobolev norm for contractive and noncontractive fixed-point operators
- Bregmanized nonlocal regularization for deconvolution and sparse reconstruction
- Comparison between \(W_2\) distance and \(\dot{H}^{-1}\) norm, and localization of Wasserstein distance
- Elliptic equations and Gaussian processes
- Elliptic Gaussian random processes
- Fractional Gaussian fields: a survey
- Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality
- Generalized inverses: theory and computations
- scientific article; zbMATH DE number 50534 (Why is no real title available?)
- scientific article; zbMATH DE number 3574917 (Why is no real title available?)
- scientific article; zbMATH DE number 1181255 (Why is no real title available?)
- scientific article; zbMATH DE number 1909499 (Why is no real title available?)
- scientific article; zbMATH DE number 936298 (Why is no real title available?)
- scientific article; zbMATH DE number 5060482 (Why is no real title available?)
- scientific article; zbMATH DE number 3199111 (Why is no real title available?)
- Image Decomposition and Restoration Using Total Variation Minimization and theH1
- Image recovery using functions of bounded variation and Sobolev spaces of negative differentiability
- Image recovery via nonlocal operators
- Image restoration and decomposition via bounded total variation and negative Hilbert-Sobolev spaces
- Image sharpening via Sobolev gradient flows
- Negative norms and boundary problems
- Non-blind and blind deconvolution under Poisson noise using fractional-order total variation
- Nonlinear total variation based noise removal algorithms
- Optimal Transport for Parameter Identification of Chaotic Dynamics via Invariant Measures
- Optimal transport with proximal splitting
- Partial differential equation methods for image inpainting
- Poisson image denoising based on fractional-order total variation
- Principles of computerized tomography imaging
- Sobolev active contours
- Stability of Gibbs posteriors from the Wasserstein loss for Bayesian full waveform inversion
- The DC (Difference of convex functions) programming and DCA revisited with DC models of real world nonconvex optimization problems
- The quadratic Wasserstein metric for inverse data matching
- The Rational SPDE Approach for Gaussian Random Fields With General Smoothness
- The Split Bregman Method for L1-Regularized Problems
- Total generalized variation
- Variational Methods in Image Processing
- Variational multiplicative noise removal by DC programming
- Variational regularisation for inverse problems with imperfect forward operators and general noise models
- Very singular diffusion equations: second and fourth order problems
- Video stabilization of atmospheric turbulence distortion
This page was built for publication: Implicit regularization effects of the Sobolev norms in image processing
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6130343)
