Model - Driven Image Analysis to Augment

In this paper we consider how information may be obtained from images. To search large image collections we need to search on secondary parameters. We may look for images containing certain types of objects, for images where the objects are of a certain size or shape, or for images having certain features. Since we now have techniques to rapidly acquire and store many images, we need techniques for automatic image analysis to generate such parameters. This paper describes a promising category of image analysis, namely model-driven methods. Two examples , operating in very diierent domains, are presented.