ABSTRACT
The segmentation of bone tissue from CT images is a complex process. It is difficult to find a solution that works in all cases. In a bone, there are two very distinct zones: cortical and cancellous tissue. Cortical tissue is very dense and it can be found in the outer part of the bone. Cancellous tissue is mainly in the inner part of the bone. This type of tissue is more heterogeneous and it has less intensity in a CT image. In addition, the intensity value for the same tissue differs between slices. This happens with both cortical and cancellous tissue. For instance, intensity values on the epiphysis and the diaphysis are different in a long bone. Near the joints, the cortical zone is very thin. This zone even disappears in the area closest to the join. Therefore, the transition of the intensity values near the joints generally appears to be fuzzy and some areas within the bone may have similar intensity than the soft tissue surrounding the bone. This may cause incomplete segmentation or overgrowing (Lee, Lai, Hu, Huang, Tsai, & Ueng 2012).
