1 Introduction
Object categorization has been an active topic of research in psychology and computer vision for decades. Initially, vision scientists and psychologists formulated hypotheses about models of object categorization and recognition [7], [8], [25]. Subsequently, in the past 10 years or so, object recognition and categorization have become very popular areas of research in computer vision. With two general models emerging, generative and discriminative, the newly developed algorithms aim to adhere to the original modeling constraints proposed by vision scientists. For example, the hypothesis put forth by Biederman et al. [1] suggests five classes of relations between an object and its setting that can characterize the organization of objects into real-world scenes. These are: (i) interposition (objects interrupt their background), (ii) support (objects tend to rest on surfaces), (iii) probability (objects tend to be found in some contexts but not others), (iv) position (given an object is probable in a scene, it often is found in some positions and not others), and (v)familiar size (objects have a limited set of size relations with other objects).