A density distance augmented Chan-Vese active contour for CT bone segmentation

Bone segmentation from Computed Tomography (CT) images is a critical component in computer-assisted orthopedic surgery but is a challenging task. Among many active contour (AC) models employed to solve the problem, the Chan-Vese AC [1] yields superior performances as evaluated in [2]. However, the CV AC fails to correctly extract the objects that are of high inhomogeneity because its nature is the minimization of the differences within the objects. In this paper, we propose to incorporate a Bhattacharrya term to the CV functional which helps to maximize the distance between the density functions of the objects and the background. The proposed model is tested with various synthetic and real CT images. Preliminary experimental results show that it can overcome the limitation of the CV AC.