Multiscale segmentation of three-dimensional MR brain images

Segmentation of MR brain images using intensity values is severely limited owing to field inhomogeneities, susceptibility artifacts and partial volume effects. Edge based segmentation methods suffer from spurious edges and ga ps in boundaries. A multiscale method to MRI brain segmentation is presente d which uses both edge and intensity information. First a multiscale repres entation of an image is created, which can be made edge dependent to favor intra-tissue diffusion over inter-tissue diffusion. Subsequently a multisca le linking model (the hyperstack) is used to group voxels into a number of objects based on intensity. It is shown that both an improvement in accurac y and a reduction in image post-processing can be achieved if edge dependen t diffusion is used instead of linear diffusion. The combination of edge de pendent diffusion and intensity based linking facilitates segmentation of g rey matter, white matter and cerebrospinal fluid with minimal user interact ion. To segment the total brain (white matter plus grey matter) morphologic al operations are applied to remove small bridges between the brain and cra nium. If the total brain is segmented, grey matter, white matter and cerebr ospinal fluid can be segmented by joining a small number of segments. Using a supervised segmentation technique and MRI simulations of a brain phantom for validation it is shown that the errors are in the order of or smaller than reported in literature.