Development of noise free hybrid segmentation approach in MRI processing

Abstract

Brain tumor is the second most common reason of the death happening in humans. The growing field of image processing consistently allows living organs and organisms to be explored non-invasively. The aim of this paper is to highlight the presence of noise in a MR image, kind of noises and their effects. The development of hybrid segmentation followed by filtering technique for high variation of noise in MR images is done. For further analysis removal of noise plays a key role in processing the information of an image. Various Filters such as Mean filter, Median Filter, Adaptive filter are examined and applied for the removal of Gaussian Noise, Salt and Pepper Noise, Speckle noise and are also compared in terms of RMSE and PSNR values.