A Nonlinear Anisotropic GVF-shock Diffusion Method to Multi-frame Image Super-resolution Reconstruction

摘要

This paper presents a new nonlinear anisotropic Gradient Vector Flow (GVF) shock diffusion method to multi-frame image Super-resolution (SR) reconstruction. Firstly, matrix-valued diffusion tensor is constructed based on anisotropic nonlinear structure tensor, making full use of its directional smooth characteristics. Then, an anisotropic GVF is proposed to diffuse along tangent and normal directions separately, combing with adaptively weighted shock filter to enhance edge sharpness. Experimental results show that proposed algorithm enhances image edges and suppresses noise effectively.