基于脉冲神经网络的红外目标提取

模拟生物信息处理机制，设计了一种用于红外目标提取的脉冲神经网络(SNN)。首先，利用输入层脉冲神经元将激励图像转化为脉冲序列；其次，采用中间层脉冲神经元输出脉冲的密度编码红外图像目标的轮廓像素和非目标轮廓像素；最后，根据输出层神经元输出脉冲的密度是否超过阈值提取红外目标。实验结果表明，设计的脉冲神经网络具有较好的红外目标提取性能，并且符合生物视觉信息处理机制。

Target extraction in infrared image based on spiking neural networks

A Spiking Neural Network (SNN) for target extraction in infrared image was designed by means of a simulation of bio-inspired information processing mechanism. Firstly, the infrared image stimulus were transferred into spike trains by neurons in input layer of the spiking neural network; then, the target outline in infrared image was specially encoded by the density of spiking trains in middle layer of the spiking neuron network; at last, the outline pixels of the infrared target was determined by whether the firing density of the corresponding neuron in the networks output layer was over a threshold. The experimental results show that the designed spiking neuron network has a good performance in infrared target extraction, and is more biologically realistic than the existing methods.