Transforms Without Tears

Abstract

In the previous chapter we have considered shape detection per se and have decided (with a little biased guidance!) that parametric transformation is a good way of partitioning points on the boundary of a shape (Edge Image Data) into sets. Each set is then labelled by the parameters associated with the shape formed by those points. In order to be able to put such a theoretically powerful technique to good use it is necessary not just to ‘know’ about it but to understand it in a way that enables it to be implemented and manipulated to best advantage. However, the process of integral transformation is not an intuitively obvious one. Difficulty in visualizing the transformation process and its results may be compounded in that the intended user may not feel happy or safe handling sophisticated mathematical concepts. It is to these people that this chapter is addressed.