A Splitting Algorithm for Medical Image Denoising

Abstract

In this work we consider a stable algorithm for integrating a mathematical model based on mean curvature motion equation proposed in (Alvarez, Lions, Morel 1992) for image denoising. The scheme is constructed using a finite difference space discretisation and semi-implicit time discretisation and is considered with a splitting algorithm that can be implemented in parallel. We apply this algorithm to the problem of denoising optical coherence tomograms from the human retina while preserving image features.