PDF version of this document
More considerable work has been put into shapes and the selection of landmarks to define point distribution models (PDM's)10.7. Work of this nature continues at the time of writing of this report, but clear goals are already known and their establishment is yet incomplete. It is inarguably desirable and even expected that improvement will be made not only to registration, but also to the correspondence problem in shapes. More importantly, revelations in one should affect work on the other. This way, neither of the two should ever lag behind the other in terms of performance or quality; implementation will not become overly forked either. Work on shapes can be sub-categorised as follows10.8.