Proximity constraints in deformable models for cortical surface identification

Abstract

Automatic computer processing of large multi-dimensional images such as those produced by magnetic resonance imaging (MRI) is greatly aided by deformable models. A general method of deforming polyhedra is presented here, with two novel features. Firstly, explicit prevention of non-simple (self-intersecting) surface geometries is provided, unlike conventional deformable models which merely discourage such behaviour. Secondly, simultaneous deformation of multiple surfaces with inter-surface proximity constraints provides a greater facility for incorporating model-based constraints into the process of image recognition. These two features are used advantageously to automatically identify the total surface of the cerebral cortical gray matter from normal human MR images, accurately locating the depths of the sulci even where under-sampling in the image obscures the visibility of sulci. A large number of individual surfaces (N=151) are created and a spatial map of the mean and standard deviation of the cerebral cortex and the thickness of cortical gray matter are generated. Ideas for further work are outlined.