Adaptive fourth-order partial differential equation filter for image denoising

Abstract

To overcome the staircasing effects and simultaneously avoid edge blurring, this paper describes a fourth-order partial differential equation based edge-preserving regularization filter for noise removal. This technique is closely related to the nonlinear anisotropic diffusion. Compared results distinctly demonstrate the superiority of our proposed scheme over the LLT model, in terms of removing noise while sharply maintaining the edge features.