ParNes: A rapidly convergent algorithm for accurate recovery of sparse and approximately sparse signals (Q372857): Difference between revisions
From MaRDI portal
Changed an Item |
Changed an Item |
||
| Property / describes a project that uses | |||
| Property / describes a project that uses: NESTA / rank | |||
Normal rank | |||
Revision as of 02:06, 29 February 2024
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | ParNes: A rapidly convergent algorithm for accurate recovery of sparse and approximately sparse signals |
scientific article |
Statements
ParNes: A rapidly convergent algorithm for accurate recovery of sparse and approximately sparse signals (English)
0 references
21 October 2013
0 references
This article develops an algorithm for underdetermined least squares problems with 1-norm constraint on the solution. It relies on a modified Nestorev method that periodically updates the prox-center optimally and renders the proposed method almost always linearly convergent. The method is then applied to find the minimum 1-norm solution to an underdetermined least squares problem when teamed with the Pareto root finder. Numerical tests show that these methods achieve comparable numerical results in comparable time for such a problem in image processing and other areas of application.
0 references
least squares problem
0 references
Nesterov's method
0 references
basis pursuit
0 references
denoising problem
0 references
Pareto root finder
0 references
convex minimization
0 references
numerical example
0 references
algorithm
0 references
1-norm constraint
0 references
minimum 1-norm solution
0 references
image processing
0 references
