Parveen, Runa; (2011) Morphological Brain Volumetry: theory, concepts and application to quantitative tissue classification. Masters thesis (M.Phil), UCL (University College London).

Preview

Text Parveen_10209410_thesis.pdf Download (43MB) | Preview

Abstract

Magnetic Resonance Imaging can provide high resolution volumetric images of the brain with superb soft tissue contrast - segmentation is then a post-processing operation to label complex structures with complicated shapes. This leads to the development of quantitative algorithms to analyze the neuro-anatomical structures. This thesis describes research which has attempted to develop, and make improvements to, image analysis methods in order to accurately classify the anatomical structures of MR brain images. Some common artefacts, such as noise, intensity inhomogeneity and partial volume effects exist in MR brain images that can be seriously detrimental to image analysis methods. This thesis presents the background, methodology, results and further work of a project that employs morphological volumetric analysis to segment brain tissues in three dimensions. The clinical motivation behind this work is the desire to assists in brain morphomtery and in measuring the effect of any abnormalities to the brain. The partial volume effect is a major obstacle to the accurate separation of tissues in MR images. This thesis presents an affinity based fuzzy segmentation to classify the brain tissues into WM, GM and CSF in order to counteract the partial volume effect, and the classification process has taken into account the spatial connectivity along with fuzzy labelling. This algorithm is tested on well-established simulated MR brain volumes to generate an extensive quantitative comparison with different noise levels and different slice thicknesses. These novel works are applied to 20 normal clinical MR brain datasets and an extensive accuracy measures are taken into account to validate this project. The intensity non-uniformity of MR brain images is also analysed in this thesis to clinical MR brain images and demonstrate the improve results over some well-established methods. The possibilities of the future work are discussed at the end.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item