[1]张勇,何泽裕,赵东宁,等.基于差分进化优化的BP神经网络图像复原方法[J].深圳大学学报理工版,2018,35(4):405-412.[doi:10.3724/SP.J.1249.2018.04405]
 ZHANG Yong,HE Zeyu,ZHAO Dongning,et al.An image restoration method for BP neural network based on differential evolution optimization[J].Journal of Shenzhen University Science and Engineering,2018,35(4):405-412.[doi:10.3724/SP.J.1249.2018.04405]
点击复制

基于差分进化优化的BP神经网络图像复原方法()
分享到:

《深圳大学学报理工版》[ISSN:1000-2618/CN:44-1401/N]

卷:
第35卷
期数:
2018年第4期
页码:
405-412
栏目:
电子与信息科学
出版日期:
2018-07-10

文章信息/Info

Title:
An image restoration method for BP neural network based on differential evolution optimization
作者:
张勇1何泽裕1赵东宁2张席3
1)深圳大学ATR国防科技重点实验室, 广东深圳518060;2)深圳大学信息工程学院, 广东深圳518060;3)深圳技术大学基础教学部, 广东深圳 518118
Author(s):
ZHANG Yong1 HE Zeyu1 ZHAO Dongning2 and ZHANG Xi3
1) ATR Key Laboratory of National Defense Technology, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China 2) College of Information Engineering, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China 3) Department of Fundamental Teaching, Shenzhen Technology University, Shenzhen 518118, Guangdong Province, P.R.China
关键词:
模糊图像图像处理图像复原BP神经网络 差分进化峰值信噪比结构相似性
Keywords:
blurred image image processing image restoration BP neural network differential evolution peak signal to noise ratio structural similarity
分类号:
TP 319
DOI:
10.3724/SP.J.1249.2018.04405
文献标志码:
A
摘要:
针对 BP (back-propagating)神经网络收敛速度慢、易陷入局部最小的问题,基于差分进化算法,改进其差分策略,提出随机缩放差分进化(random scaling-differential evolution, RSDE)优化的BP神经网络(RSDE-BP)图像复原方法.该方法用高斯噪声对无噪图像进行模糊处理,将模糊图像和原图像组成训练对,用于训练和优化RSDE-BP算法.最后利用训练好的BP神经网络对测试图像进行复原,从而达到去除噪声的目的.仿真结果表明,与 BP神经网络、PSO-BP算法和DE-BP算法相比,所提出的算法收敛速度快,迭代次数少,且复原图像在峰值信噪比和结构相似性等指标方面有很好效果.与自适应全变差复原方法和二阶广义全变差正则项复原方法相比,该方法能够较好地恢复被噪声和模糊污染的图像,同时可以很好地保留图像的纹理和细节信息.
Abstract:
Aiming at the problems of slow convergence and local minimum of the back-propagating (BP) neural network, an image restoration method based on random scaling-differential evolution (RSDE) is proposed. In the proposed method, the noise-free images are blurred?by the Gaussian noise. The blurred images and the noise-free images are set to the training pairs, which are used to train and optimize the proposed RSDE-BP method. Finally, the optimized BP neural network is utilized to restore the test images and remove the noises. The experimental results show that the convergence rate of the RSDE-BP algorithm is faster and the number of iterations is less than the BP method, the PSO-BP method and DE-BP method. In addition, the peak signal to noise ratio (PSNR) and the structural similarity (SSIM) are better. Compared with the adaptive total variation (ATV) image deblurring method and the blurred image restoration method based on the second-order total generalized variation regularization (TGV), the RSDE-BP method can effectively restore the images polluted by the noise and blur, while preserving the image texture and details more effective.

相似文献/References:

[1]张 敏,阮双琛,杨 珺,等.连续太赫兹波实时透射成像实验研究[J].深圳大学学报理工版,2007,24(4):384.
 ZHANG Min,RUAN Shuang-chen,YANG Jun,et al.Experimental study of continuous-wave terahertz radiation real-time transmission imaging[J].Journal of Shenzhen University Science and Engineering,2007,24(4):384.
[2]胡涛,郭宝平,郭轩.基于游程分析轮廓提取算法的改进[J].深圳大学学报理工版,2009,26(4):405.
 HU Tao,GUO Bao-ping,and GUO Xuan.An improved run-based boundary extraction algorithm[J].Journal of Shenzhen University Science and Engineering,2009,26(4):405.
[3]胡媛媛,牛夏牧.基于视觉阈值的结构相似度图像质量评价算法[J].深圳大学学报理工版,2010,27(2):185.
 HU Yuan-yuan and NIU Xia-mu.Image quality assessment based on human visibility threshold theory and structural similarity[J].Journal of Shenzhen University Science and Engineering,2010,27(4):185.
[4]宋远佳,张炜,杨正伟,等.固体火箭发动机壳体脱黏缺陷的热波检测[J].深圳大学学报理工版,2012,29(No.3(189-282)):252.[doi:10.3724/SP.J.1249.2012.03252]
 SONG Yuan-jia,ZHANG Wei,YANG Zheng-wei,et al.Debond defect detection in shell of solid rocket motor by thermal wave nondestructive testing[J].Journal of Shenzhen University Science and Engineering,2012,29(4):252.[doi:10.3724/SP.J.1249.2012.03252]
[5]黄宗福,孙刚,陈曾平. 大视场空间目标光电探测起伏背景抑制算法[J].深圳大学学报理工版,2012,29(No.6(471-580)):471.[doi:10.3724/SP.J.1249.2012.06471]
 HUANG Zong-fu,SUN Gang,and CHEN Zeng-ping.A background clutter suppression algorithm for space target detection in wide field-of-view opto-electronic observation[J].Journal of Shenzhen University Science and Engineering,2012,29(4):471.[doi:10.3724/SP.J.1249.2012.06471]
[6]吴庆阳,曾祥军,黄锦辉,等.数字印模口内三维扫描技术研究[J].深圳大学学报理工版,2013,30(No.1(001-110)):60.[doi:10.3724/SP.J.1249.2013.01060]
 Wu Qingyang,Zeng Xiangjun,Huang Jinhui,et al.Study on digital impression for intraoral 3D scanning[J].Journal of Shenzhen University Science and Engineering,2013,30(4):60.[doi:10.3724/SP.J.1249.2013.01060]
[7]张敏,权润爱,苏红,等.光泵连续太赫兹波在生物成像中的应用研究(英文)[J].深圳大学学报理工版,2014,31(2):160.[doi:10.3724/SP.J.1249.2014.02160]
 Zhang Min,Quan Runai,Su Hong,et al.Investigation of optically pumped continuous terahertz laser in biological imaging[J].Journal of Shenzhen University Science and Engineering,2014,31(4):160.[doi:10.3724/SP.J.1249.2014.02160]
[8]李霞,李富生,陈园琴.基于视觉灵敏度与DCT系数的显著性检测[J].深圳大学学报理工版,2014,31(5):464.[doi:10.3724/SP.J.1249.2014.05464]
 Li Xia,Li Fusheng,and Chen Yuanqin.Saliency detection model based on human visual sensitivity and DCT coefficients[J].Journal of Shenzhen University Science and Engineering,2014,31(4):464.[doi:10.3724/SP.J.1249.2014.05464]
[9]李璟,倪东,李胜利,等.超声图像中胎儿头围的自动测量[J].深圳大学学报理工版,2014,31(5):455.[doi:10.3724/SP.J.1249.2014.05455]
 Li Jing,Ni Dong,Li Shengli,et al.The automatic ultrasound measurement of fetal head circumference[J].Journal of Shenzhen University Science and Engineering,2014,31(4):455.[doi:10.3724/SP.J.1249.2014.05455]
[10]邱文胜,牛丽红,苏秉华,等.基于ARM的嵌入式超分辨率复原系统设计[J].深圳大学学报理工版,2015,32(3):311.[doi:10.3724/SP.J.1249.2015.0]
 Qiu Wensheng,Niu Lihong,Su Binghua,et al.Design of embedded super-resolution restoration system based on ARM[J].Journal of Shenzhen University Science and Engineering,2015,32(4):311.[doi:10.3724/SP.J.1249.2015.0]

更新日期/Last Update: 2018-06-20