Low-cost denoising and deblurring using a novel nonlinear diffusion technique

An algorithm for the treatment of images affected by both blurring and salt&pepper noise is proposed with a cost only proportional to the number of pixels. The methodology uses an ad-hoc discretization scheme for the Laplace operator, multiplied by a suitable nonlinear term depending on the gradient. Even if this approach resembles a diffusion type algorithm, only one step of the procedure is in general needed, leading to significant time savings. The procedure is successfully tested on some standard black&white images.