A variational model based on split Bregman method for multiplicative noise removal

摘要

It is well known that total variation (TV) regularizer leads to the staircase effect, the higher order variational methods give rise to the restored image blurred. In this paper, we propose a novel variational model for multiplicative noise removal. The proposed model can automatically adjust the first and second order regularization terms. To solve such an objective function effectively, the split Bregman and primal-dual methods are employed in our numerical algorithm. Our experimental results show that the proposed method is more effective to filter out the multiplicative noise compared with the recent methods. (C) 2015 Elsevier GmbH. All rights reserved.