Diffusion tensor imaging has been used extensively in the brain for identifying morphological characteristics based on gray and white matter structures; however, similar applications in the spinal cord have not been as successful. Spatial resolution and limited contrast have undermined accurate delineation between gray and white matter in the spinal cord. This study applied fuzzy logic to diffusion tensor images of the human spinal cord to discriminate between gray and white matter. The technique used common anisotropy indices and newly developed indices based on properties of the diffusion ellipsoid. Preliminary applications to subjects with varying levels of spinal cord injury are also presented in this study. Results indicated larger contrast between gray and white matter compared to the traditional fractional anisotropy index and thus presents a superior technique for determining precise morphology of the spinal cord.