An improved approach for denoising MRI using non local means filter

Abstract

Various types of noise affect Magnetic Resonance Images and consequently interrupt in correct diagnosis. Thus, noise reduction is a major task while working with Magnetic Resonance (MR) Images. The process of denoising reduces the undesirable noise, but at the same time the denoising process needs to preserve the image features as well. Various techniques are available for noise removal from MR Images. Non local means filtering approach is a popular denoising technique for MR images. Modifications have been performed on the original Non Local Means algorithm to apply it on numerous applications. The aim of this paper is to denoise MRI by a novel approach, making use of a median and wiener filter in conjunction with the Non local means filtering technique.