A new steplength selection for scaled gradient methods with application to image deblurring
From MaRDI portal
Publication:897119
DOI10.1007/S10915-015-9991-9zbMATH Open1328.65138arXiv1407.2375OpenAlexW2030583950MaRDI QIDQ897119FDOQ897119
Authors: F. Porta, M. Prato, Luca Zanni 
Publication date: 17 December 2015
Published in: Journal of Scientific Computing (Search for Journal in Brave)
Abstract: Gradient methods are frequently used in large scale image deblurring problems since they avoid the onerous computation of the Hessian matrix of the objective function. Second order information is typically sought by a clever choice of the steplength parameter defining the descent direction, as in the case of the well-known Barzilai and Borwein rules. In a recent paper, a strategy for the steplength selection approximating the inverse of some eigenvalues of the Hessian matrix has been proposed for gradient methods applied to unconstrained minimization problems. In the quadratic case, this approach is based on a Lanczos process applied every m iterations to the matrix of the most recent m back gradients but the idea can be extended to a general objective function. In this paper we extend this rule to the case of scaled gradient projection methods applied to non-negatively constrained minimization problems, and we test the effectiveness of the proposed strategy in image deblurring problems in both the presence and the absence of an explicit edge-preserving regularization term.
Full work available at URL: https://arxiv.org/abs/1407.2375
Recommendations
- A scaled gradient projection method for constrained image deblurring
- Nonnegative least-squares image deblurring: improved gradient projection approaches
- Faster gradient descent and the efficient recovery of images
- Scaling techniques for gradient projection-type methods in astronomical image deblurring
- Gradient projection algorithm for total variation image restoration by adaptive steplength selection rules
Numerical mathematical programming methods (65K05) Large-scale problems in mathematical programming (90C06) Computing methodologies for image processing (68U10) Numerical methods for inverse problems for integral equations (65R32)
Cites Work
- Nonlinear total variation based noise removal algorithms
- Title not available (Why is that?)
- Title not available (Why is that?)
- Rank-Deficient and Discrete Ill-Posed Problems
- Convex optimization theory.
- Two-Point Step Size Gradient Methods
- An algorithm for total variation minimization and applications
- A first-order primal-dual algorithm for convex problems with applications to imaging
- A discrepancy principle for Poisson data
- Title not available (Why is that?)
- A scaled gradient projection method for constrained image deblurring
- A Nonmonotone Line Search Technique for Newton’s Method
- Computational Methods for Inverse Problems
- Analysis of bounded variation penalty methods for ill-posed problems
- This is SPIRAL-TAP: Sparse Poisson Intensity Reconstruction ALgorithms—Theory and Practice
- New adaptive stepsize selections in gradient methods
- On spectral properties of steepest descent methods
- Deblurring Images
- Alternate minimization gradient method
- An efficient gradient method using the Yuan steplength
- An Interior Trust Region Approach for Nonlinear Minimization Subject to Bounds
- On the filtering effect of iterative regularization algorithms for discrete inverse problems
- A new stepsize for the steepest descent method
- Duality-based algorithms for total-variation-regularized image restoration
- A limited memory steepest descent method
- An alternating extragradient method for total variation-based image restoration from Poisson data
- Gradient methods with adaptive step-sizes
- Inexact spectral projected gradient methods on convex sets
- Regularization parameter selection methods for ill-posed Poisson maximum likelihood estimation
- Penalized maximum likelihood image restoration with positivity constraints: multiplicative algorithms
- Sparse Poisson Noisy Image Deblurring
- On the convergence of primal-dual hybrid gradient algorithms for total variation image restoration
- Scaling techniques for gradient projection-type methods in astronomical image deblurring
- Efficient gradient projection methods for edge-preserving removal of Poisson noise
- A modified projection algorithm for large strictly-convex quadratic programs
- A general method to devise maximum-likelihood signal restoration multiplicative algorithms with non-negativity constraints.
- Nonnegative image reconstruction from sparse Fourier data: a new deconvolution algorithm
- Accelerated gradient methods for the x-ray imaging of solar flares
- A convergent blind deconvolution method for post-adaptive-optics astronomical imaging
Cited In (17)
- A comparison of edge-preserving approaches for differential interference contrast microscopy
- Hybrid limited memory gradient projection methods for box-constrained optimization problems
- Nonnegative least-squares image deblurring: improved gradient projection approaches
- Delayed Gradient Methods for Symmetric and Positive Definite Linear Systems
- Steplength selection in gradient projection methods for box-constrained quadratic programs
- On Quasi-Newton Forward-Backward Splitting: Proximal Calculus and Convergence
- Comparison of active-set and gradient projection-based algorithms for box-constrained quadratic programming
- Using gradient directions to get global convergence of Newton-type methods
- Variable metric techniques for forward-backward methods in imaging
- On the steplength selection in gradient methods for unconstrained optimization
- Reconstruction of 3D X-ray CT images from reduced sampling by a scaled gradient projection algorithm
- A VMiPG method for composite optimization with nonsmooth term having no closed-form proximal mapping
- Convergence of Inexact Forward--Backward Algorithms Using the Forward--Backward Envelope
- A scaled gradient projection method for constrained image deblurring
- A variable metric forward-backward method with extrapolation
- Dirichlet problem for a nonlocal \(p\)-Laplacian elliptic equation
- A new proximal heavy ball inexact line-search algorithm
Uses Software
This page was built for publication: A new steplength selection for scaled gradient methods with application to image deblurring
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q897119)
