Transformation of 3-D object for object segmentation in 3-D medical image

摘要

The present invention defines a method and apparatus for segmenting an object from a three-dimensional image. First, a surface of the object is estimated using scale estimation processes such as Laplacian-of-Gaussian filters in one, two or three dimensions. Secondly, the estimated surface is transformed by having a shrink function applied to it independently in each dimension to give a regularly shaped (preferably substantially spherical), non-isotropic shape. The object is segmented from the volume around it and inverse-transformed to give the original shape, which is then isolated from the image volume.