Optimization issues in medical imaging and fiber-tracking

Abstract

Fiber tracking is a relatively recent methodology, made possible by access to new highly advanced MR scanners able to produce high-quality diffusion tensor images (DTI), which promises clinicians a possibility to observe the actual trajectories of fibers and the connectivity of the living human brain. Analysis of such data provides clinicians with a unique tool for diagnosing and predicting disease which affect the “wiring” of the brain, such as stroke, tumors, multiple sclerosis and Alzheimer’s disease. The purpose of this PhD thesis has been to explore ways to improve the quality of DTI as well as the accuracy of the fiber tractography which can be estimated from such data. In order to achieve this purpose we have investigated two fundamentally different areas to improve DTI results; data pre-processing and more accurate data descriptors.