Denoising Magnetic Resonance Imaging Using Fuzzy Similarity Based Filter

Abstract

Images are very useful source of information which is often degraded due to presence of noise. Noise present in the image especially in MRI images hides the important information which is very important to diagnose the disease. So to retain the quality of image we need to remove noise. Hence denoising is very essential to obtain precise images to facilitate the accurate observations. Fuzzy Similarity based Non-Local Means (FSNLM) filter is used to select homogeneous pixels for the estimation of noise-free pixels. Rician noise introduces bias which corrupts MRI images. The bias correction has been proposed for the removal of bias from MRI images which increases contrast and PSNR. The proposed scheme has been tested on simulated data sets and compared with existing method.