Application of Denoising Non-Local Mean Filter (NLM) in MRI Brain Image T2WI TSE SENSE

Abstract

The acquisition of MRI images takes a long time. One of the efforts to improve the examination acquisition time is by using a parallel imaging technique, namely SENSE. However, the SENSE technique has a weak-point that is reducing SNR. Reducing SNR induces an increase in the amount of noise. One of denoising method that is able to increase SNR is Non-local mean filter (NLM). Denoising at post-image acquisition is a cheaper and more effective alternative. This research is expected to produce faster scanning times and maintain the quality of MRI images. The purpose of this study is to analyze the effectiveness of denoising Nonlocal Mean Filter (NLM) on MRI brain image information T2WI TSE SENSE on r factor. Experimental research on MRI brain images with T2WI TSE sequences use parallel imaging SENSE application. R factor that is used in the application of SENSE is 2. The improvement of the image information is executed by using denoising Non-Local Mean (NLM) technique. The image evaluation is carried out by comparing pre- and post-using denoising qualitatively (brain anatomical information). The results of the parametric test show that there is a difference in anatomical information on MRI Brain T2WI TSE images by using parallel imaging SENSE between pre- and post-using denoising Non-Local Means (NLM) technique with p-value <0.001 (< 0, 05). Nonlocal Mean Filter (NLM) effectively enhances anatomic MRI brain information T2WI TSE SENSE r factor 2.