In order to represent we need some model of our objects. Since we sense surface information with visual process and this is highly dependent on the geometry of the object we consider aspects of how we model such data. Note for some visual processes we may need to consider reflectance properties of the surface also -- we do not consider such matters here.
We will survey some of the major representation scheme which are not necessarily restricted to Vision or AI subjects. Indeed the topics discussed below have are subjects of study in their own right.
Some AI related representation schemes have been developed. For example Semantic networks we have already dealt with this subject and we will leave its application to vision until later (see Model Based Object Recognition)