Massively parallel optical fuzzy-neuro-chaos system for processing information depicted by pictures

摘要

A new optical system capable of understanding the meaning of a picture is proposed. The picture transforms and conveys instantaneously the information of the painter's imagery to viewers. This system can be formed using massively parallel optical fuzzy-neural networks that can recognize geometric patterns on the basis of incoherent optics. When the system applies the output to the input and the reference, the system can gaze and pursue a particular figure. In this system, the light intensity of the output image becomes chaotic. This change of the light intensity represents the characteristic of a thinking machine that behaves as the appreciator diverts his visual point, and interest in various parts of the picture according to a change of impression. The system understands the meaning of the picture on the basis of synthesizing impressions. The paper shows the fundamental optical diagram of this system and demonstrates that the system was able to gaze and pursue a particular figure.