For the resulting PDEs from these high order models only explicit or semi-implicit time-marching methods have been reported in the literature. One of the advantages of our SFP1 and SFP2 algorithms is that they can be very easily generalized to solve other PDEs similar to (2.2). That is, we can implement fast fixed point algorithms for other models by splitting their differential operators and adding up suitable stabilizing terms. Then these fixed point algorithms can be used as smoothers in an MG context. For instance, we already have successfully applied this idea to solving the fourth order PDE of the Euler’s elastica digital inpainting model; see [11] for reported results. We also have encouraging results from using this method for solving a three-dimensional denoising problem also known as surface fairing that has been studied by Elsey and Esedog ̄lu [26]. For this we have implemented only the two-dimensional case or curve denoising.