Distinguishing Computer-Generated Images from Natural Images Using Channel and Pixel Correlation

doi:10.1007/s11390-020-0216-9

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 三维可视计算

	Distinguishing Computer-Generated Images from Natural Images Using Channel and Pixel Correlation
	Zhang, Rui-Song 3,4; Quan, Wei-Ze 3,4; Fan, Lu-Bin 2; Hu, Li-Ming 1; Yan, Dong-Ming 1,3,4
刊名	JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY
	2020-05-01
卷号	35 期号:3 页码:592-602
关键词	natural image computer-generated image channel and pixel correlation convolutional neural network
ISSN号	1000-9000
DOI	10.1007/s11390-020-0216-9
通讯作者	Yan, Dong-Ming(yandongming@gmail.com)
英文摘要	With the recent tremendous advances of computer graphics rendering and image editing technologies, computergenerated fake images, which in general do not reflect what happens in the reality, can now easily deceive the inspection of human visual system. In this work, we propose a convolutional neural network (CNN)-based model to distinguish computergenerated (CG) images from natural images (NIs) with channel and pixel correlation. The key component of the proposed CNN architecture is a self-coding module that takes the color images as input to extract the correlation between color channels explicitly. Unlike previous approaches that directly apply CNN to solve this problem, we consider the generality of the network (or subnetwork), i.e., the newly introduced hybrid correlation module can be directly combined with existing CNN models for enhancing the discrimination capacity of original networks. Experimental results demonstrate that the proposed network outperforms state-of-the-art methods in terms of classification performance. We also show that the newly introduced hybrid correlation module can improve the classification accuracy of different CNN architectures.
资助项目	National Key Research and Development Program of China[2019YFB2204104] ; Beijing Natural Science Foundation of China[L182059] ; National Natural Science Foundation of China[61772523] ; National Natural Science Foundation of China[61620106003] ; National Natural Science Foundation of China[61802406] ; Alibaba Group through Alibaba Innovative Research Program ; Joint Open Research Fund Program of State Key Laboratory of Hydroscience and Engineering and Tsinghua-Ningxia Yinchuan Joint Institute of Internet of Waters on Digital Water Governance
WOS关键词	GRAPHICS ; IDENTIFICATION ; CLASSIFICATION
WOS研究方向	Computer Science
语种	英语
出版者	SCIENCE PRESS
WOS记录号	WOS:000539025300009
资助机构	National Key Research and Development Program of China ; Beijing Natural Science Foundation of China ; National Natural Science Foundation of China ; Alibaba Group through Alibaba Innovative Research Program ; Joint Open Research Fund Program of State Key Laboratory of Hydroscience and Engineering and Tsinghua-Ningxia Yinchuan Joint Institute of Internet of Waters on Digital Water Governance
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/39799]
专题	模式识别国家重点实验室_三维可视计算
通讯作者	Yan, Dong-Ming
作者单位	1.Tsinghua Univ, State Key Lab Hydrosci & Engn, Beijing 100084, Peoples R China 2.Alibaba Grp, Hangzhou 310023, Peoples R China 3.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China 4.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Rui-Song,Quan, Wei-Ze,Fan, Lu-Bin,et al. Distinguishing Computer-Generated Images from Natural Images Using Channel and Pixel Correlation[J]. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY,2020,35(3):592-602.
APA	Zhang, Rui-Song,Quan, Wei-Ze,Fan, Lu-Bin,Hu, Li-Ming,&Yan, Dong-Ming.(2020).Distinguishing Computer-Generated Images from Natural Images Using Channel and Pixel Correlation.JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY,35(3),592-602.
MLA	Zhang, Rui-Song,et al."Distinguishing Computer-Generated Images from Natural Images Using Channel and Pixel Correlation".JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY 35.3(2020):592-602.