Image Parallel Processing Based On Anisotropic Diffusion Equations

The modeling of spatial discontinuities for problems such as surface recovery, image denoising and segmentation has been intensely studied in computer vision during recent years. In this paper, we present a class of equation model for image denoising and edge enhancement. The model is based on robust statistics error norm and variational theory. By analysing the parameters of the equation and designing the methods of module partitions, the numerical experiments reach a good result and the parallel efficiency has been improved remarkably with finite difference method. According to the image characteristics, the parallel algorithms based on finite element method is constructed. Because of the advantages of the cutting methods, the total rigidity matrix is symmetric and positive definite. So, conjugate gradient method is used to improve the parallel efficiency. By improving the methods of module partitions, the matrix storages is reduced and the parallel efficiency is improved remarkably. The numerical experiments get better results and the method can be applicated extensively relatively.