Second‐order PDE‐based image restoration algorithm using directional diffusion

Variational and partial differential equation (PDE)‐based algorithms have been widely applied in image restoration. However, it may produce undesirable staircase effect or blur image edges. To avoid these problems, a second‐order PDE model based on directional diffusion has been proposed for image restoration. This model can just diffuse along the edge's tangential direction of the original image. Thus it can preserve the edges, avoid the staircase effect in the restored image. Visual and quantitative results demonstrate that the proposed second‐order PDE model is superior to other models in preserving edges and avoiding staircase effect for image restoration.