A Nonlocal Model for Reconstructing Images Corrupted by Cauchy Noise

The aim of this paper is to present the mathematical and numerical study of a nonlocal nonlinear model based on the variable exponent p(x)-Laplacian for removing Cauchy noise, which is a type of impulsive and non-Gaussian degradation. The proposed model benefits from the performance of the nonlocal approach to preserve small details and textures, and the efficiency of the variable exponent to reduce the execution time. To demonstrate the reliability of our proposed model, we provide some experimental denoising results and illustrate the comparison with some models from the literature.