DMCA
Parameter Selection for Total Variation Based Image Restoration Using Discrepancy Principle
| Citations: | 5 - 2 self |
BibTeX
@MISC{Wen_parameterselection,
author = {You-wei Wen and Raymond H. Chan},
title = { Parameter Selection for Total Variation Based Image Restoration Using Discrepancy Principle},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The key issues in solving image restoration problem successfully are: the estimation of the regularization parameter which balances the data-fidelity with the regularity of the solution; and the development of efficient numerical techniques for computing the solution. In this paper, we derive a fast algorithm that simultaneously estimates the regularization parameter and restores the image. The new approach is based on total-variation (TV) regularized strategy and Morozov discrepancy principle. The TV norm is represented by the dual formulation that changes the minimization problem into a minimax problem. A proximal point method is developed to compute the saddle point of the minimax problem. By adjusting the regularization parameter adaptively in each iteration, the solution is guaranteed to satisfy the discrepancy principle. We will give the convergence proof of our algorithm and show numerically that it is better than some state-of-the-art methods in both speed and accuracy.
Keyphrases
total variation image restoration using discrepancy principle parameter selection regularization parameter minimax problem discrepancy principle key issue fast algorithm dual formulation minimization problem tv norm morozov discrepancy principle convergence proof new approach state-of-the-art method saddle point image restoration problem efficient numerical technique proximal point method
