基于广义回归神经网络的图像修复算法
|
摘要
根据人类视觉感知的连通原理,提出一种基于广义回归神经网络的图像修复算法,该算法通过对图像数据进行回归分析确定缺失区域,具有简单和有效的特点。根据尺寸对缺失区域进行分离以及分类,将广义回归神经网络用于每个缺失区域,以便修复损坏的像素点。在文字去除、划痕去除以及噪声去除3个应用方面,对该算法的性能进行评估,采用客观衡量方式对修复图像的视觉质量进行评估。实验结果表明了该算法的有效性和可靠性。
A new image restoration algorithm based on generalized regression neural network was proposed according to the human visual perception.It was simple and effective to determine the missing area by regression analysis of the image data.According to the size,the missing area was separated and classified.The generalized regression neural network was used for each missing area,to repair the damaged pixels.The performance of the proposed algorithm was evaluated in three applications including word removal,scratch removal and noise removal,and the visual quality of the restored image was evaluated using the objective measurement.Experimental results show that the proposed algorithm is effective and reliable.
A new image restoration algorithm based on generalized regression neural network was proposed according to the human visual perception.It was simple and effective to determine the missing area by regression analysis of the image data.According to the size,the missing area was separated and classified.The generalized regression neural network was used for each missing area,to repair the damaged pixels.The performance of the proposed algorithm was evaluated in three applications including word removal,scratch removal and noise removal,and the visual quality of the restored image was evaluated using the objective measurement.Experimental results show that the proposed algorithm is effective and reliable.
引文
[1]ZHAI Donghai,YU Jiang,DUAN Weixia,et al.Image restoration of damaged zone partition[J].Chinese Journal of Image and Graphics,2014,19(6):835-842(in Chinese).[翟东海,鱼江,段维夏,等.破损区域分块划分的图像修复[J].中国图象图形学报,2014,19(6):835-842.]
[2]MA Shuang,TAN Yuanpeng,XU Gang.Image restoration[J].Computer Aided Design and Computer Graphics Aided Design and Computer Graphics,2015,27(2):271-278(in Chinese).[马爽,谈元鹏,许刚.块关联匹配与低秩矩阵超分辨融合的图像修复[J].计算机辅助设计与图形学学报,2015,27(2):271-278.]
[3]HAN Mingzhu,GUO Shuxu,ZANG Lingling,et al.An improved algorithm for image restoration based on sample block[J].Computer Engineering and Application,2015,51(16):192-194(in Chinese).[韩明珠,郭树旭,臧玲玲,等.基于样本块的图像修复改进算法[J].计算机工程与应用,2015,51(16):192-194.]
[4]HU Wenjin,LIU Zhongmin,LI Zhanming.An improved wavelet domain image restoration algorithm[J].Computer Science,2014,41(5):299-303(in Chinese).[胡文瑾,刘仲民,李战明.一种改进的小波域图像修复算法[J].计算机科学,2014,41(5):299-303.]
[5]HE Wenxi,YE Kuntao.Digital image restoration based on BP neural network[J].Journal of Jiangxi University of Science and Technology,2014,35(1):65-69(in Chinese).[贺文熙,叶坤涛.基于BP神经网络的数字图像修复[J].江西理工大学学报,2014,35(1):65-69.]
[6]Dash R,Majhi B.Motion blur parameters estimation for image restoration[J].Optik-International Journal for Light and Electron Optics,2014,125(5):1634-1640.
[7]Chung B,Yim C.Hybrid error concealment method combining exemplar-based image inpainting and spatial interpolation[J].Signal Processing Image Communication,2014,29(10):1121-1137.
[8]Xie J,Xu L,Chen E,et al.Image denoising and inpainting with deep neural networks[J].Advances in Neural Information Processing Systems,2012,32(1):350-358.
[9]Zhang K,Gao X,Tao D,et al.Single image super-resolution with non-local means and steering kernel regression[J].IEEE Transactions on Image Processing a Publication of the IEEE Signal Processing Society,2012,21(11):4544-4556.
[10]Jung YM,Shen JJ.Illusory shapes via first-order phase transition and approximation[J].Journal of Mathematical Imaging&Vision,2015,53(3):303-313.
[11]Wang X,Wang L,Pujin LU.Research on color image restoration based on harmonic neural network[J].Computer Engineering&Applications,2015,45(1):488-492.
[2]MA Shuang,TAN Yuanpeng,XU Gang.Image restoration[J].Computer Aided Design and Computer Graphics Aided Design and Computer Graphics,2015,27(2):271-278(in Chinese).[马爽,谈元鹏,许刚.块关联匹配与低秩矩阵超分辨融合的图像修复[J].计算机辅助设计与图形学学报,2015,27(2):271-278.]
[3]HAN Mingzhu,GUO Shuxu,ZANG Lingling,et al.An improved algorithm for image restoration based on sample block[J].Computer Engineering and Application,2015,51(16):192-194(in Chinese).[韩明珠,郭树旭,臧玲玲,等.基于样本块的图像修复改进算法[J].计算机工程与应用,2015,51(16):192-194.]
[4]HU Wenjin,LIU Zhongmin,LI Zhanming.An improved wavelet domain image restoration algorithm[J].Computer Science,2014,41(5):299-303(in Chinese).[胡文瑾,刘仲民,李战明.一种改进的小波域图像修复算法[J].计算机科学,2014,41(5):299-303.]
[5]HE Wenxi,YE Kuntao.Digital image restoration based on BP neural network[J].Journal of Jiangxi University of Science and Technology,2014,35(1):65-69(in Chinese).[贺文熙,叶坤涛.基于BP神经网络的数字图像修复[J].江西理工大学学报,2014,35(1):65-69.]
[6]Dash R,Majhi B.Motion blur parameters estimation for image restoration[J].Optik-International Journal for Light and Electron Optics,2014,125(5):1634-1640.
[7]Chung B,Yim C.Hybrid error concealment method combining exemplar-based image inpainting and spatial interpolation[J].Signal Processing Image Communication,2014,29(10):1121-1137.
[8]Xie J,Xu L,Chen E,et al.Image denoising and inpainting with deep neural networks[J].Advances in Neural Information Processing Systems,2012,32(1):350-358.
[9]Zhang K,Gao X,Tao D,et al.Single image super-resolution with non-local means and steering kernel regression[J].IEEE Transactions on Image Processing a Publication of the IEEE Signal Processing Society,2012,21(11):4544-4556.
[10]Jung YM,Shen JJ.Illusory shapes via first-order phase transition and approximation[J].Journal of Mathematical Imaging&Vision,2015,53(3):303-313.
[11]Wang X,Wang L,Pujin LU.Research on color image restoration based on harmonic neural network[J].Computer Engineering&Applications,2015,45(1):488-492.