Image retrieval based on object's orientation spatial relationship

Similarity-based retrieval of images is an important task in many image dat abase applications. A major class of users' requests requires retrieving th ose images in the database that are spatially similar to the query image. W e propose an approach for computing the orientation spatial similarity betw een two symbolic images in this paper. The proposed approach is not only ro tation invariant, but also captures the relative distance and orientation r ange between objects. (C) 2001 Elsevier Science B.V. All rights reserved.