This thesis presents some image processing tools for plastic surgery planning. In particular,
it presents a novel method that combines local and global context in a probabilistic
relaxation framework to identify cephalometric landmarks used in Maxillofacial plastic
surgery. It also uses a method that utilises global and local symmetry to identify abnormalities
in CT frontal images of the human body. The proposed methodologies are
evaluated with the help of several clinical data supplied by collaborating plastic surgeons.