ABSTRACT
The hypothesis of invariants in the sequence of perspective transformations was implicit. The formless invariants of the face of an object specify that particular face of the object. The hypothesis of invariants, even in its early and less radical form, has met with a certain incredulity from perceptionists and artists. The information for perception is in the form, not in any formless invariants, and, in this case, object-perception does depend on form perception. There are a number of facts and lines of reasoning to support the hypothesis that invariants are information for perception, even in the special case of pictorially mediated perception. The information-bearing features are things like the following: alignment or straightness as against bentness or curvature; perpendicularity or rectangularity; parallelity as against convergence; intersections; closures and symmetries.
