Abstract

Digital optical cellular arrays are single-instruction-multiple-data (SIMD) arrays of many low complexity (fine-grain) processing elements (PEs). The PEs themselves can be implemented by electronic or optoelectronic methods. These arrays have many general applications in numerical processing and symbolic substitution computing. They are particularly suited to bit plane images (images in which each PE represents a pixel, and each pixel takes on the value 0 or 1). In this application, each PE is referred to as a cell and is responsible for computing the output of one image pixel according to a single instruction broadcast to all PEs from a central control unit.

© 1991 Optical Society of America