Model-based segmentation of CT images

This paper presents preliminary work on the segmentation of Computed Tomography data using a model-based approach. Conventional image processing of CT data is aimed at the production of simple iso-surfaces for surgical planning or diagnosis — such methods are not suitable for the automated detection of fractures, which is the ultimate application of our work. To address these deficiencies a surface-based technique with appropriate constraints is introduced. The output of the segmentation phase is a triangulated surface representing the bone or bones of interest. We illustrate the method applied to low resolution CT test data and discuss its robustness and performance.