Color restoration of mural image based on double constrained convolutional neural network

doi:10.13245/j.hust.200602

	Color restoration of mural image based on double constrained convolutional neural network
	Xu, Zhigang 2; Yin, Wenyu 2; Zhu, Xufeng 1
刊名	Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition)
	2020-06-23
卷号	48 期号:6 页码:6-12
关键词	Color Convolution Convolutional neural networks Discoloration Markov processes Multilayer neural networks Restoration Color features Color restoration Difference constraints Global feature Higher-order Image-based Local structure Markov Random Fields
ISSN号	16714512
DOI	10.13245/j.hust.200602
英文摘要	Aiming at the problem of discoloration in grotto mural images,a color restoration method for mural images based on double constrained convolutional neural network was proposed.A convolutional neural network was used to extract the higher-order features of the mural image.Based on this,in order to carry out global restoration of mural color,the maximum mean difference constraint was used to establish a global feature mapping relationship to restore the global color feature in mural image.For the lack of local color information in different convolutional layers,Markov random field constraint was used to restrict the spatial layout.This constraint can retain the local structure information of the mural image and restore the local color of mural image.In addition,the square gradient constraint was used to suppress the noise in the image restoration process.The experimental results show that this method is better in both subjective and objective indicators and has a better overall restoration effect in the faded or discolored areas of mural images. © 2020, Editorial Board of Journal of Huazhong University of Science and Technology. All right reserved.
语种	中文
出版者	Huazhong University of Science and Technology
内容类型	期刊论文
源URL	[http://ir.lut.edu.cn/handle/2XXMBERH/115491]
专题	计算机与通信学院
作者单位	1.Beijing Aerospace Automatic Control Institute, Beijing; 100854, China 2.School of Computer and Communication, Lanzhou University of Technology, Lanzhou; 730050, China;
推荐引用方式 GB/T 7714	Xu, Zhigang,Yin, Wenyu,Zhu, Xufeng. Color restoration of mural image based on double constrained convolutional neural network[J]. Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition),2020,48(6):6-12.
APA	Xu, Zhigang,Yin, Wenyu,&Zhu, Xufeng.(2020).Color restoration of mural image based on double constrained convolutional neural network.Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition),48(6),6-12.
MLA	Xu, Zhigang,et al."Color restoration of mural image based on double constrained convolutional neural network".Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) 48.6(2020):6-12.