Gliomas are the most common primary brain tumor in adults, causing a high mortality rate. The World Health Organization classified gliomas into four grades according to their histological features. Low-grade gliomas are not benign but grow slowly for several years, and portend better prognosis; however, high-grade gliomas are malignant and grow fast. The grading is currently performed by visual investigation and histopathologic diagnosis, but there are intra-observer variability and sampling error. Several image-based grading methods have been proposed and investigated; the diffusion-weighted magnetic resonance imaging technique is one of the most promising methods for the grading of gliomas. Diffusion-weighted magnetic resonance images can measure the diffusion of water molecules in the tissue, therefore diffusion-weighted magnetic resonance images can represent the microstructure of the tissue. In this thesis, we study various methods for glioma grading on diffusion-weighted magnetic resonance images including reduced field-of-view diffusion-weighted magnetic resonance images.