Image denoising is one of the initial stages of image processing. Many models based on the diffusion method have been used to smooth the image. One of the problems, we face in the model based on the diffusion method is its possible loss of edges. The diffusion force is known to be more effective in areas of high frequency. So This paper suggests combining the direction of isophote and fourth-order partial differential equations to reduce the problem of loss edges and preserve important details of the image. The direction of the isophote can regulate the direction of diffusion. Thus, we have a proposed model that can remove the noise in the area while preserving the important edges and details of the image. We have proven the efficiency and superiority of the proposed model by applying it to a set of images and solving it numerically using the finite difference method (FDM).
Read full abstract