Non-local Means Denoising using An Adaptive Kernel

Non-local means algorithm is one of the powerful image denoising methods. Maintaining noise near edges and textural parts of a noisy image, is one of the main drawbacks ofNLM. In this paper we introduce an adaptive kernel derived from image structure to remove maintained noise. Experimentalresults show superiority of our algorithm in comparison with original NLM as well as a method based on shape adaptive patches.

Keywords:

Non-local means , Denoising , Shape adaptive patches , Adaptive kernel