Background: Accurate extraction of bone contours from two-dimensional (2D) projective X-ray images is an important component for computer-assisted diagnosis, planning or three-dimensional (3D) reconstruction.
Methods: We propose a 3D statistical model-based, fully automatic segmentation framework for extracting the proximal femur contours from calibrated X-ray images. The automatic initialization is an estimation of a Bayesian network algorithm to fit a multiple-component geometrical model to the X-ray data. The contour extraction is accomplished by a non-rigid 2D/3D registration between the statistical model and the X-ray images, in which bone contours are extracted by a graphical model-based Bayesian inference.
Results: The contour extraction algorithm was verified on both cadaver and clinical datasets, visually and quantitatively. Compared to the 'gold standard', a mean error of 1.6 mm was observed when the automatically extracted contours were used to reconstruct a patient-specific surface model.
Conclusions: Our statistical model-based bone contour extraction approach holds the potential to facilitate the application of 2D/3D reconstruction in surgical navigation.