Segmentation and classification of brain mri using deformable, statistical And knowledge based approaches with Fft techniques

Abstract

Magnetic resonance imaging MRI is a powerful medical imaging newlinemodality to produce high resolution images with good contrast of the newlinedifferent biological soft tissue types As a non invasive technique the large newlinequantity of data provided by MRI for brain imaging in particular aids newlinestatisticians and medical professionals in disease diagnosis and functional newlineunderstanding of the human brain In particular interest to this study is the newlineclassification of three main tissue types in the brain Gray Matter GM newlineWhite Matter WM and CerebroSpinal Fluid CSF and the segmentation of newlinemagnetic resonance images in patients with gliomas newlinePathological regions from MRI scans of normal adults and patients newlinewith neuro degenerative diseases are needed for improved understanding of newlinedisease progression in vivo As images are often confounded by acquisition newlinenoise and partial volume effects developing an automatic robust and newlineefficient segmentation is essential to the accurate quantification of disease newlineseverity A major challenge of this work is to devise robust techniques to newlineaddress the above said problems newline newline