Variational approachin image restoration problems

Abstract

Image restoration has been an active research area in image processing and computer vision during the past several decades. We explore variational partial differential equations (PDE) models in image restoration problem. We start our discussion by reviewing classical models, by which the works of this dissertation are highly motivated. The content of the dissertation is divided into two main subjects. First topic is on image denoising, where we propose non-convex hybrid total variation model, and then we apply iterative reweighted algorithm to solve the proposed model. Second topic is on image decomposition, in which we separate an image into structural component and oscillatory component using local gradient constraint.