A regularized nonlinear diffusion approach for texture image denoising

Abstract

In this paper a new partial differential equation based method is presented with a view to denoising images having textures. The proposed model combines a nonlinear anisotropic diffusion filter with recent harmonic analysis techniques. A wave atom shrinkage allied to detection by gradient technique is used to guide the diffusion process so as to smooth and maintain essential image characteristics. Two forcing terms are used to maintain and improve edges, boundaries and oscillatory features of an image having irregular details and texture. Experimental results show the performance of our model for texture preserving denoising when compared to recent methods in literature. © 2009 IEEE.