基于生物机制脉冲神经网络的特征提取

Feature Extraction Based on Biological Mechanism Spiking Neural Network

脉冲神经元可以被用于处理生物刺激并且可以解释大脑复杂的智能行为。脉冲神经网络以非常逼近生物的神经元模型作为处理单元,可以直接用来仿真脑科学中发现的神经网络计算模型,输出的脉冲信号还可与生物神经系统对接。而小波变换是一个非常有利的时频分析工具,它可以有效的压缩图像并且提取图像的特征。本文中将提出一种与人类视觉系统的开/关神经元阵列相结合的脉冲神经网络,来实现针对视觉图像的快速小波变换。仿真结果显示,这个脉冲神经网络可以很好地保留视觉图像的关键特征。

The human visual system has the ability to selectively attend to certain locations while ignoring others in a typical complexity of the visual environment.The functionalities of spiking neurons can be applied to deal with biological stimuli and explain complicated intelligent behaviors of the brain.Visual images are transferred among these neurons in human visual system in the form of spiking trains through the ON or OFF pathways.This paper try to simulate how the human brain uses volition-controlled method to extracts useful image information.Wavelet transform is a powerful time-frequency analysis tool that can efficiently compress image and extract image features.In this article,a simplified conductancebased integrate-and-fire spiking neural network model combined with the ON/OFF neuron arrays associated with the human visual system is proposed to perform the fast wavelet transform for visual images.The simulation results show that the spiking neural network can preserve the key features of visual images very well.