An optical neural network used for content addressable two-dimensional associative memory
作者
Yanping Sun
出处
期刊:International Conference on Optoelectronic Science and Engineering '90日期:2017-12-04卷期号:: 259-259
标识
DOI:10.1117/12.2294897
摘要
An optical neural network system used for content addressable 2-D associative memory has been demonstrated. A liquid crystal light valve (LCLV) is employed as a 2-D parallel thresholding and feedback device. The interconnections of the network are realized with a specially prepared hologram. When one of the stored images, even if only less than half of the image, appears at the input plane, the system can recognize and give a complete retrieval of the input image.