A Novel Partial Differential Equation Method Based on Image Features and Its Applications in Denoising

Abstract

In this paper a novel denoising algorithm based on partial differential equation variational approach is proposed. The proposed algorithm is obtained by weighted combinations of the an isotropic diffusion model and total variation model. Experimental results show that our new approach is more efficient in image denoising than both an isotropic diffusion and total variation in terms of preserving edges information and texture features.