OME methods and results have been discussed in the section about experiments. That section dealt with two distinct families of problems, but the main one was image registration and the relations it has to models of appearance. From this point onwards in this chapter, discussions will concentrate on that one later portion of the work. For the realisation of one successful approach and for completeness, this section will also explain the way in which registration is performed at present.
A later part of this chapter outlines the structure of the successful approach graphically. It is followed by the proposal of a two new approaches to be attempted in the near future if time permits. These more illustrative parts may be hard to follow, but references should clear up the way to full understanding. They should also serve as a starting point for greater exploration of this approach and fore-coming plans as they currently stand.