A kernel-based technique for image enhancement based on anisotropic diffusion model

Image noise reduction is a fundamental operation in image processing. The current work considers an image noise reduction approach based on an anisotropic diffusion process. An efficient localized computational technique based on the radial kernel functions, RBF-FD, is used to deal with the governing nonlinear diffusion model. A domain decomposition framework has been implemented to reduce the computational cost of the process. Some numerical experiments are presented to show the effectiveness of the proposed method.