White matter hyperintensities are lesions in the brain that exhibit high signal intensity on T2-weighted FLAIR images. In this study, we designed a seed-based segmentation method of these lesions from T2-weighted images of 1.3mm axial resolution, utilizing the local gradient field on a breadth-first search scheme as well as level thresholding and smoothing for false-positive reduction. Additionally, since clinical MRIs are at lower resolution of 1x1x5mm, we modeled the relationship between lesion cross-sectional radius and its true volume with a cubic polynomial fit to correct for stereological bias.
This abstract and the presentation materials are available to members only; a login is required.