Structure-preserved NLTV regularization for image denoising

Abstract

Image denoising is an important problem in image processing since noise may interfere with visual or automatic interpretation. This paper proposes a novel Nonlocal Total Variation (NLTV) regularization method to reduce noise in digital images. The data fidelity term in variational framework of NLTV is implemented via iterative nonlocal means, which can preserve the structure information in a denoised image. Experimental results show that our method is very competitive with the NLTV method, especially in preserving image structure and introducing very few artifacts.