The Iterative Regularization Method Based on Image Decomposition Model

Motivated by iterative regularization procedure based on total variation, we develop concretely a new iterative regularization procedure for Vese-Osher image decomposition model. The new algorithm improves significantly the quality of restoration when Vese-Osher standard model is used to real textured image denoising. The new algorithm is very simple, and we implement the method by the semi-implicit finite differences.