Image restoration using total variation and anisotropic diffusion equation

Abstract

This paper proposes a new model for the image restoration which combines the total variation minimization with the pure  anisotropic diffusion equation of Alvarez and Morel. According to the introduction of new diffusion term, this model can not only remove noise but also enhance edges and keep their locality. And it can also keep textures and large-scale fine features that are not characterized by edges. Due to these favorable characteristics, the processed images turn much clearer and smoother, meanwhile, their significant details are kept, which results in appealing vision.

Keywords

total variation, anisotropic diffusion equation, noise removal, detail subject classification