Skip to Main Content
The studying models of complex diffusion PDEs have potential meaning in representing a sophisticated process of real world and have a lot of use in practice. In this paper we propose a two layer CNN - 2D structure for image edge enhancing and denoising by complex nonlinear diffusion. In the first part of this paper, complex PDEs and CNN's architectures are briefly presented. In the second part the complex linear diffusion is considered. In this part we also presented a CNN model that is used for speeding up the solving of complex linear diffusion PDE in real time. In the third part, an image processing method for enhancing ramp edge and denoising by complex nonlinear diffusion CNN is given. Simulations are demonstrated in part 4 and finally some conclusions and future works are summarized.