An adaptive-focus statistical shape model for segmentation and shape modeling of 3-D brain structures

This paper presents a deformable model for automatically segmenting brain s tructures from volumetric magnetic resonance (MR) images and obtaining poin t correspondences, using geometric and statistical information in a hierarc hical scheme. Geometric information is embedded into the model via a set of affine-invariant attribute vectors, each of which characterizes the geomet ric structure around a point of the model from a local to a global scale, T he attribute vectors, in conjunction with the deformation mechanism of the model, warranty that the model not only deforms to nearby edges, as is cust omary in most deformable surface models, but also that it determines point correspondences based on geometric similarity at different scales. The prop osed model is adaptive in that it initially focuses on the most reliable st ructures of interest, and gradually shifts focus to other structures as tho se become closer to their respective targets and, therefore, more reliable. The proposed techniques have been used to segment boundaries of the ventri cles, the caudate nucleus, and the lenticular nucleus from volumetric MR im ages.