A statistical approach to image denoising based on kriging and PODs with comments on physical versus statistical models
Abstract
Physical models are based on a priori physical information. Typically conservation laws, kinematic descriptions, constitutive laws, boundary conditions, i.e., partial differential equations which are then numerically integrated with solutions in the volume and/or at the boundary. They aim at being predictive with little experimental data.