Volume morphing for segmentation of bone from 3D data
Johanna Pettersson, Magnus Borga, Mats Andersson, Hans E. Knutsson · 2005
This paper introduces a method for automatic segmentation of bones in 3D data. Morphons [4] are employed for the segmentation process. This method is based on an iterative scheme where a prototype model, the Morphon, describing the general shape of the target structure is registered to the input data according to accumulation of certainty weighted displacement fields. The estimation of the displacement field is based on minimizing the difference in quadrature filter phase between the input data and the prototype data. The method has shown promising results for automatic segmentation of the femoral and the pelvic bone from volumetric computed tomography (CT) data, where thin structures, such as the joint space and weak contours, often causes troubles in the segmentation process. 1