A 3D statistical shape model of the pelvic bone for segmentation

Abstract

Statistical models of shape are a promising approach for robust and automatic segmentation of medical image data. This work describes the construction of a statistical shape model of the pelvic bone. An interactive approach is proposed for solving the correspondence problem which is able to handle shapes of arbitrary topology, suitable for the genus 3 surface of the pelvic bone. Moreover it allows to specify corresponding anatomical features as boundary constraints to the matching process. The model's capability for segmentation was tested on a set of 23 CT data sets. Quantitative results will be presented, showing that the model is well suited for segmentation purposes.