8.1. Introduction   

Computer Scientists have long sought to find the most efficient and effective way of representing images in terms of some form of data structure. The aim is to find an image representation technique that is both efficient in space and access time whilst not losing any accuracy of the original image - whether the image is two dimensional or the scene is three dimensional.

A computerised representation of a two dimensional image is usually in terms of a collection of "pixels". The attributes of this image are usually traceable back to the nature of how this atomic unit of measurement - the pixel, is shaped and how it relates to other pixels in the image (Bell et al,1988). One of the most researched and significant data structures for two dimensinal image representation is the quadtree. Due to its naturally hierachical form, its implementation blends itself effectively to representing the original image. For three dimensional images, octrees will be examined as a data structure to store the three dimensional graphical scene. Like the quadtree data structure, octrees are hierarchically based and represent the image at the object level (instead of pixel).