Abstract

Associative memories and associative processing have many applications in symbolic and parallel computing.¹ Optical schemes, in particular those based on holographic principles, are well suited to associative processing because of their high parallelism and storage capacity. Previous workers² have demonstrated that holographically stored images can be recalled using relatively complicated reference images, but did not utilize nonlinear feedback to reduce the cross-talk which results when multiple images are stored and a partial or distorted input is used in addressing the memory. These earlier approaches were limited in their ability to reconstruct an image from a set of partial input images.

© 1986 Optical Society of America