Learning spatial relationships in computer vision

Spatial relationships exhibited among regions in an image play an important role in the interpretation of a scene. While humans have an innate ability to recognize spatial relations, it has been difficult to produce algorithms to model these relationships. There have been several attempts at defining spatial relationships between regions in a digital image, most recently, with the use of fuzzy set theory. In a previous paper, we compared three algorithmic methods for defining spatial relations to gain insight into this complex situation. Here, we examine the ability of neural network structures along with fuzzy integration to generalize spatial relationship membership functions from simple examples