A Comparative Analysis of Filters for Noise Reduction and Smoothening of Brain MRI Images

Abstract

Image pre-processing is used to improve the quality of image by removing irrelevant and unwanted parts of the image such as noise. Medical images generally suffer from noise, and this noisy effect is due to the inherent properties of the medical devices such as magnetic field generated by the scanners or may be due to the vibrations and loud sound generated during Magnetic Resonance Image (MRI) scan. MRI images suffer due to radiation and physical movements of magnetic parts of the devices which are responsible for image noise in MRI. The MRI device uses strong magnetic field to create images and hence images usually suffer from different types of noises such as Gaussian, Salt and pepper, speckle etc. Removing such noise from brain MRI images is a big challenge for researchers to have better visualization of the images to accurately identify brain tumor. The core objective of this paper is to discuss different methods used in pre-processing of the brain MRI images and compare the quality of the resultant images generated by different pre-processing techniques. This paper discusses different types of noises and filters suitable for these noises by comparing the histograms of various filters.