Logic programming's applications in knowledge based systems will require heavy use of images for effective human interfaces, and future graphic work stations need efficient and easily usable methods of describing images. Logic programming methods provide a promising way for naive users to manipulate images. This promise is demonstrated, and a novel application of infinite logic terms is explored. These terms arise from the lack of an occur check in the fast unification algorithms of most logic programming systems. Infinite terms are not part of traditional logic but are useful in image processing, particularly in describing recursive images. An implementation is briefly described.