p-Laplace diffusion for distance function estimation, optimal transport approximation, and image enhancement

In this paper, we propose ADMM-based numerical schemes for power-law diffusion equations involving the p-Laplacian div(|∇u|p−2∇u) with constant (p=const) and variable (p=p(x)) exponents. Applications to distance function and optimal transport approximations (p is large and constant) and image enhancement (p is small and variable) are considered.