摘要
将传统Fourier神经网络与忆阻器相结合,用忆阻器做突触,构建新型的忆阻Fourier神经网络.推导忆导变化与权值更新的关系,提出忆阻突触权值更新规则,构建单输入忆阻Fourier神经网络,提出忆阻BP算法对模糊二值图像和灰度进行处理.Matlab仿真实验表明该算法可以有效实现图像复原,提高图像清晰度.忆阻Fourier神经网络有望用于解决复杂的图像处理问题.
A new memristive Fourier neural network was proposed based on the combination of the tradi tional Fourier neural network and a memristor, a new circuit element, synapse. The relationship between memristive conductance and synapse the weight updating rule of the memristive synapse was proposed. Th with the memristor acting as the weight updating was derived, and e single input memristive Fourier neural network was built, and a memristive BP algorithm was used to process blurred binary and gray ima ges. Matlab simulation confirmed that the memristive BP algorithm proposed could realize image restora tion and heighten the legibility of images. The memristive Fourier neural network could be used for sol ving complex image processing problems.
出处
《西南大学学报(自然科学版)》
CAS
CSCD
北大核心
2014年第1期1-6,共6页
Journal of Southwest University(Natural Science Edition)
基金
新世纪优秀人才支持计划(教技函[2013]47号)
国家自然科学基金资助项目(61372139
61101233
60972155)
教育部"春晖计划"科研项目(Z2011148)
留学人员科技活动项目择优资助经费资助项目(国家级
优秀类
渝人社办[2012]186号)
重庆市高等学校优秀人才支持计划(渝教人[2011]65号)
重庆市高等学校青年骨干教师资助计划(渝教人[2011]65号)
中央高校基本科研业务费专项资金(XDJK2012A007
XDJK2013B011)
关键词
忆阻器
傅立叶神经网络
BP算法
图像复原
memristor
Fourier neural network
BP algorithm~ image restoration