论文部分内容阅读
在研究洗牌网局部互连光学实现问题的基础上 ,采用洗牌型图样间联想模型建立了具有完整的加权互连、求和、非线性处理及反馈功能的光电混合神经网络实验系统 ,进行了 8× 8数字样本的光学联想识别。实验结果证实了洗牌网理论的可行性和洗牌型图样间联想模型光学实现的优越性
On the basis of studying the optical interconnection problem in the local interconnection of shuffle network, an experimental system of optoelectronic hybrid neural network with complete interconnection, summation, nonlinear processing and feedback function is established by using shuffling pattern association model. Optical Association Recognition of 8 × 8 Digital Samples. The experimental results confirm the feasibility of the shuffling network theory and the superiority of the optical realization of the shuffled inter-pattern model