Scaled projected-directions methods with application to transmission tomography
From MaRDI portal
Publication:2218907
Abstract: Statistical image reconstruction in X-Ray computed tomography yields large-scale regularized linear least-squares problems with nonnegativity bounds, where the memory footprint of the operator is a concern. Discretizing images in cylindrical coordinates results in significant memory savings, and allows parallel operator-vector products without on-the-fly computation of the operator, without necessarily decreasing image quality. However, it deteriorates the conditioning of the operator. We improve the Hessian conditioning by way of a block-circulant scaling operator and we propose a strategy to handle nondiagonal scaling in the context of projected-directions methods for bound-constrained problems. We describe our implementation of the scaling strategy using two algorithms: TRON, a trust-region method with exact second derivatives, and L-BFGS-B, a linesearch method with a limited-memory quasi-Newton Hessian approximation. We compare our approach with one where a change of variable is made in the problem. On two reconstruction problems, our approach converges faster than the change of variable approach, and achieves much tighter accuracy in terms of optimality residual than a first-order method.
Recommendations
- A scaled gradient method for digital tomographic image reconstruction
- Tomography: problems and multiscale solutions
- Tomographic image reconstruction based on generalized projections
- A computation for tomographic reconstruction from projections
- scientific article; zbMATH DE number 723989
- Tomography with a finite set of projections: singular value decomposition and resolution
- Tomographic reconstruction from arbitrary directions using ridge functions
- Linearization of the diffraction tomography problem through construction of a scattering matrix
- A spectral projected gradient optimization for binary tomography
- Compton scattering tomography in translational geometries
Cites work
- A Limited Memory Algorithm for Bound Constrained Optimization
- A first-order primal-dual algorithm for convex problems with applications to imaging
- A local update strategy for iterative reconstruction from projections
- A projected Newton-CG method for nonnegative astronomical image deblurring
- A scaled gradient projection method for constrained image deblurring
- Algorithm 778: L-BFGS-B
- Benchmarking optimization software with performance profiles.
- CUTEst: a constrained and unconstrained testing environment with safe threads for mathematical optimization
- Fundamentals of Computerized Tomography
- Implementation of an optimal first-order method for strongly convex total variation regularization
- Large-scale active-set box-constrained optimization method with spectral projected gradients
- Line search algorithms with guaranteed sufficient decrease
- Methods of conjugate gradients for solving linear systems
- Newton's Method for Large Bound-Constrained Optimization Problems
- Projected Newton Methods for Optimization Problems with Simple Constraints
- Reconstruction of 3D X-ray CT images from reduced sampling by a scaled gradient projection algorithm
- Remark on ``Algorithm 778: L-BFGS-B: Fortran subroutines for large-scale bound constrained optimization
- Representations of quasi-Newton matrices and their use in limited memory methods
- Scaling techniques for gradient projection-type methods in astronomical image deblurring
- Testing a Class of Methods for Solving Minimization Problems with Simple Bounds on the Variables
- Two-Point Step Size Gradient Methods
- Updating Quasi-Newton Matrices with Limited Storage
Cited in
(2)
This page was built for publication: Scaled projected-directions methods with application to transmission tomography
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2218907)
