Image denoising using modified nonlinear diffusion approach

Partial Differential Equation (PDE) based, non-linear diffusion approaches are an effective way to denoise the images. In this paper, the work is extended to include anisotropic diffusion, where the diffusivity is a tensor valued function, which can be adapted to local edge orientation. This allows smoothing along the edges, but not perpendicular to it. The diffusion tensor is a function of differential structure of the evolving image itself. Such a feedback leads to nonlinear diffusion filters. It shows improved performance in the presence of noise. The original anisotropic diffusion algorithm updates each point based on four nearest-neighbor differences, the progress of diffusion results in improved edges. In the proposed method the edges are better preserved because diffusion is controlled by the gray level differences of diagonal neighbors in addition to 4 nearest neighbors using coupled PDF formulation. The proposed algorithm gives excellent results for MRI images, Biomedical images and Fingerprint images with noise.