| Malignant gliomas are a disease that affects a significant number of individuals and result in poor prognosis. Much of the research in image analysis has been qualitative, which highlights the need for objective quantitative information to provide an aid for the neuro-oncologist in treatment planning. The first chapter of the research work presented in this dissertation involves state of the art automated and semi-automated techniques for segmentation of the tumor components. The second chapter describes potential image-derived features for prognostic analysis of the tumor components and the utility of quantitative analysis. The third chapter discusses potential content-based image retrieval methods that provide an intuitive way of indexing brain tumor images and having the ability to distinguish between grade three and grade four tumors. |
