Nonlocal means algorithm using superformula kernel for image denoising

Abstract

Using the superformula, a mathematic function describing many complex shapes and curves, this paper designs a new superformula kernel (SFK). We then introduce a novel nonlocal means (NLM) algorithm for image denoising by replacing the Gaussian kernel with the SFK. Simulations and comparisons demonstrate that the proposed kernel and algorithm show excellent denoising performance in terms of the peak signal and ratio (PSNR) and structural similarity (SSIM).