论文部分内容阅读
基于细胞神经网络(CNN)细胞单元的等效电路及其电学特性模型,利用SET-MOS混合结构反相器实现了模型中的激活函数电路,用耦合电容单元实现CNN细胞的系统模板,构建了SET-MOSCNN细胞硬件电路,并将其应用在图像处理中.仿真结果表明,所设计的CNN硬件电路具有结构简单、功耗低、响应速度快等特点,可用于构成各种规模的CNN电路,进一步满足大规模信号处理的需求及提高集成电路的集成度.
Based on the equivalent circuit and the electrical characteristic model of CNN cell unit, the activation function circuit in the model was realized by using the SET-MOS hybrid structure inverter, and the CNN cell system template was realized by the coupling capacitor unit SET-MOSCNN cell hardware circuit and its application in image processing.The simulation results show that the designed CNN hardware circuit has the characteristics of simple structure, low power consumption and fast response, and can be used to form CNN circuits of various scales, Further meet the needs of large-scale signal processing and improve the integration of integrated circuits.