A Fast Algorithm for Variational Image Inpainting

In this paper, we research a class variational image inpainting models with total variation regularization. Using a splitting technique, an iterative procedure of alternately solving a pair of easy subproblems is constructed. The proposed approach has fast speed than state-of-the-art methods which need to calculate Euler-Lagrange equation. The experiments show that our algorithm visually can obtain a natural and believable inpainting results for filling in missing or damaged regions of an image.