Statistical Approach to Compare Image Denoising Techniques in Medical MR Images

Abstract

In medical image processing, magnetic resonance (MR) imaging techniques play an important role. The images acquired are usually affected from various noise such as gaussian noise, salt and pepper noise, speckle noise, periodic noise etc. Therefore, acquisition of images without noise is nearly impossible task. Various filtering techniques are used to reduce the noise for further analysis of medical images. In this paper, numerous digital filters such as mean filter, median filter, wiener filter, and adaptive filter are applied for the removal of different level of Gaussian noise, salt and pepper noise and speckle noise that are added separately in MR images. The performance of all the filtering techniques are compared on the basis of the statistical parameters such as Peak Signal to Noise Ratio (PSNR) and Root Mean Square Error (RMSE).