对脉冲耦合神经网络中被动神经元的脉冲周期分析

Pulse Cycle Analysis of Passive Neurons in the Pulse Coupled Neural Network

研究了离散PCNN中被动神经元的被动脉冲周期,首先定义了动态比较比,而不是逻辑比较来描述神经内状态与动态阈值之间的线性差异;然后利用动态比较比的最大下限,给出了一个近似准确的被动脉冲周期公式,并通过对估计和实际被动脉冲周期的误差分析,证明了该公式的合理性;此外,我们从估计的脉冲周期中推导出一个稳定的脉冲周期,从而使神经元可以连续地在两个不同的时间阶段进行非周期性和周期性的脉冲;此外,还估算了被动神经元开始周期性脉冲的初始阶段,并举例说明,结果与理论分析一致。

In order to solve the passive pulse period of passive neurons in discrete PCNN.First,the dynamic comparison ratio is defined to describe the linear difference between the inner state of the nerve and the dynamic threshold is defined.Then,using the maximum and minimum of the dynamic comparison ratio,an approximate and accurate formula for the passive pulse period is given.In addition,we derive a stable pulse period from the estimated pulse period,so that the neuron can continuously carry out non-periodic and periodic pulses in two different time stages.In addition,the initial phase of the periodic pulse of the passive neuron is estimated.