Rendering recognizably unique textures

We present a perceptually designed hardware-accelerated algorithm for generating unique background textures for distinguishing documents. To be recognizable, the texture should produce a random feature vector in the brain after visual feature extraction. Our motivating example is a hypertext user interface which shows a fragment of a link's target in the margin. Upon traversing the link, the fragment expands to fill the screen. Our goal is to avoid user disorientation by texturing each document with a unique background so that a document can easily be recognized from a fragment. The user should be able to learn the textures of the most often visited documents, as per Zipf's law. The results of an initial experiment show that the generated textures are indeed recognizable. We discuss a method for enhancing text readability by both providing fast, interactive zooming and unnoticeably bleaching the background around text.