An anisotropic diffusion PDE for noise reduction and thin edge preservation

In this paper, an image denoising operator defined by a nonlinear partial differential equation (PDE) is presented. Similar to the operators proposed by Perona, Malik and Catte et al., it can remove noise, enhance step-like edges and keep the locality of the edges. Its extra ability is to keep thin edges. The criterion for stopping time is also investigated. The new operator is capable of removing rather high uniform noise without sacrificing the details of the image. If the noise is Gaussian with not too high standard deviation, the result is also quite good