Dammann convolution optical computer

• Main Authors: ZHOU CHANGHE, YU JUNJIE, MIAO CHAOFENG, YAN YU
• Format: Patent
• Language: Chinese; English
• Published: 17.11.2020
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; OPTICAL COMPUTING DEVICES; PHYSICS

A Dammann convolution optical computer is formed by connecting n neural networks in series, n is a positive integer, each neural network comprises an input part, a convolution layer part and an outputpart, and the convolution process of the method is completed by a pure optical information processing structure. The computer has the advantages of high efficiency, large capacity, cascadability, good spatial expansibility and the like, gives full play to the parallel computing capability of optics, and is an important development direction of future optical computing. 一个达曼卷积光计算机，由n个神经网络串联构成，n为正整数，每一神经网络包括输入部分、卷积层部分和输出部分，本发明的卷积过程由纯光学信息处理结构完成。本发明具有高效率、大容量、可级联性和良好的空间拓展性等优点，充分发挥光学的并行计算能力，是未来光计算的重要发展方向。