Two-dimensional arrangement of spatial patterns representing numerical data in input images for effective use of hardware resources in digital optical computing system based on optical array logic

Abstract

The authors have proposed a parallel optical computation technique making effective use of available hardware resources without decreasing the processing efficiency in optical array logic (OAL), a paradigm for parallel digital optical computing allowing arbitrary neighborhood logical operations. No methodology for effectively arranging numerical data in the input image in digital optical computing has been reported, and so the parallel algorithms in OAL proposed in past studies required large-scale hardware for their implementation. The proposed technique segments the data to be processed and regularly arranges them in several rows in the input images. The validity of the technique has been demonstrated by applying it to two parallel operations. The authors have verified the advantages of the proposed technique in terms of processing efficiency compared to the conventional technique.