Identification of Iron Deposition in Brain Using Magnetic Resonance Imaging

Abstract

Iron deposition in brain has been observed with normal aging and is associated with neurodegenerative diseases. The accumulation of iron in brain is found to be the major cause for various neurodegenerative diseases Alzheimer s disease, Parkinson s disease, Multiple Sclerosis and Huntington s disease. The automated classification of Magnetic Resonance Images on the basis of iron deposition in the basal ganglia region of brain has not yet been performed. It is very difficult to analyse iron regions in brain using simple Magnetic Resonance Imaging techniques. The Magnetic Resonance Imaging sequence namely Susceptibility Weighted Imaging helps to distinguish the brain iron regions. newlineThe objective of this research work is to investigate the iron regions in selected areas of basal ganglia region of brain and classify the Magnetic Resonance Images based on iron deposition. The Magnetic Resonance Images are usually prone to noise like rician and gaussian noise. It is very difficult to perform image processing functions with the presence of noise. Initially the Magnetic Resonance Image is pre-processed using the enhanced model of gaussian smoothing called as improved gaussian smoothing technique. An attempt has been made in this work to investigate the best method for denoising the Magnetic Resonance Images. The improved Gaussian smoothing technique for denoising has been compared with various traditional filters. newlineAfter performing image pre-processing, the region of interest is selected based on the areas of interest. In this research, the identification of iron content in the selected areas of basal ganglia are only considered. These areas of interest include caudate, putamen and globus pallidus. First, the region of interest is marked on the input images through manual ROI selection with the help of an expert radiologist. Patches are identified by selecting core pixels from various areas of the input image. These pixels forms the center pixels for the patches.