帮助 关于我们

返回检索结果

多色彩通道特征融合的GAN合成图像检测方法
GAN Synthetic Image Detection Using Fused Features in the Multi-Color Channels

查看参考文献45篇

乔通 1,2,3   陈彧星 1   谢世闯 1   姚恒 4   罗向阳 3 *  
文摘 当前,生成对抗网络(Generative Adversarial Networks, GAN)合成的逼真图像难以识别,严重危害国家网络安全及社会稳定.与此同时,多数基于深度神经网络模型设计的检测器需要大规模训练样本,且存在模型可解释度不高、泛化性能差等问题.为了克服上述亟待解决的关键性难题,本文提出一种多色彩通道特征融合的GAN合成图像检测方法.首先,探索分析真实自然图像和GAN合成图像在不同色彩空间相邻像素之间的差异,并设计差异度量算法,完成色彩通道选择.其次,利用图像像素间的高度相关性,在八个方向上通过二阶马尔可夫链对相邻像素之间的差分数组进行建模,提取差分像素邻接矩阵特征.最后,利用上述特征,设计一种简单且高效的集成分类器完成GAN合成图像的检测任务.在基于StyleGAN模型合成的伪造人脸数据集中,所提出方法的检测准确率高达100.00%;在小样本训练约束条件下,正负样本对数仅仅为2时,检测准确率高达99.65%;在单类样本训练约束条件下,正样本数仅仅为50时,检测准确率高达92.84%.在基于更先进的StyleGAN2和PGGAN模型合成的伪造场景数据集中,所提出方法的检测准确率达到99.96%以上.以上大量实验表明,本文所提出的方法明显优于比较的GAN合成图像检测方法.本文方法已经开源:https://github. com/cyxcyx559/ccss.
其他语种文摘 Currently, it is very difficult to identify the images synthesized by generative adversarial networks (GAN), which severely poses the threat on national cyber security and social stability. Meanwhile, most classifiers based on deep neural networks require large-scale samples for training, where the problems such as low model interpretability and poor generalization performance are less addressed. To overcome the limitations, we propose to design the ensemble classifier using fused features in the multi-color channels. First of all, by studying the discrimination of adjacent pixels in the multi-color channels between natural and GAN synthetic images, the difference metric is designed based on the correlation of adjacent pixels, in order to select the optimal color channels. Secondly, by utilizing the highly-correlated relationship among pixels, the difference array between adjacent pixels are modeled through a second-order Markov chain along eight directions, and meanwhile the subtractive pixel adjacency matrix features are successfully extracted. Finally, based on the extracted features, a simple but efficient detector for identifying GAN synthetic images is constructed. In the image dataset synthesized by the StyleGAN model, the results show that the accuracy of the proposed detector can reach 100.00%. It can also identify GAN synthetic images very well when the pair number of positive and negative training samples is 2 (99.65% accuracy) or only 50 positive training samples are provided (92.84% accuracy). The accuracy can also reach more than 99.96% in the image dataset synthesized by StyleGAN2 and PGGAN models. Numerous experiments show that the proposed method in this paper is better than the compared forensic methods. Our code is available at https://github.com/cyxcyx559/ccss.
来源 电子学报 ,2024,52(3):924-936 【核心库】
DOI 10.12263/DZXB.20220711
关键词 图像取证 ; 色彩通道 ; 特征融合 ; 生成对抗网络 ; 马尔可夫链 ; 集成分类器
地址

1. 杭州电子科技大学网络空间安全学院, 浙江, 杭州, 310016  

2. 中国科学院信息工程研究所, 信息安全国家重点实验室, 北京, 100093  

3. 河南省网络空间态势感知重点实验室, 河南省网络空间态势感知重点实验室, 河南, 郑州, 450001  

4. 上海理工大学光电信息与计算机工程学院, 上海, 200093

语种 中文
文献类型 研究性论文
ISSN 0372-2112
学科 自动化技术、计算机技术
基金 浙江省属高校基本科研业务费专项资金 ;  中国科学院自动化研究所模式识别国家重点实验室开放基金 ;  河南省网络空间态势感知重点实验室开放课题基金 ;  国家重点研发计划 ;  国家自然科学基金 ;  中原科技创新领军人才项目
文献收藏号 CSCD:7706814

参考文献 共 45 共3页

1.  梁瑞刚. 视听觉深度伪造检测技术研究综述. 信息安全学报,2020,5(2):1-17 CSCD被引 10 次    
2.  Goodfellow I. Generative adversarial networks. Communications of the ACM,2020,63(11):139-144 CSCD被引 786 次    
3.  Gu J X. Recent advances in convolutional neural networks. Pattern Recognition,2018,77:354-377 CSCD被引 266 次    
4.  Pevny T. Steganalysis by subtractive pixel adjacency matrix. Proceedings of the 11th ACM workshop on Multimedia and security,2009:75-84 CSCD被引 1 次    
5.  Denton E. Deep generative image models using a Laplacian pyramid of adversarial networks. Proceedings of the 28th International Conference on Neural Information Processing Systems-Volume 1,2015:1486-1494 CSCD被引 2 次    
6.  Zhang H. StackGAN: Text to photorealistic image synthesis with stacked generative adversarial networks. 2017 IEEE International Conference on Computer Vision (ICCV),2017:5907-5915 CSCD被引 1 次    
7.  Zhang H. StackGAN++: Realistic image synthesis with stacked generative adversarial networks. IEEE Transactions on Pattern Analysis and Machine Intelligence,2019,41(8):1947-1962 CSCD被引 61 次    
8.  Karras T. Progressive growing of GANs for improved quality, stability, and variation,2017 CSCD被引 10 次    
9.  Gulrajani I. Improved training of wasserstein GANs. Proceedings of the 31st International Conference on Neural Information Processing Systems,2017:5769-5779 CSCD被引 73 次    
10.  Zhang Z Z. Photographic text-toimage synthesis with a hierarchically-nested adversarial network. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,2018:6199-6208 CSCD被引 2 次    
11.  Brock A. Large scale GAN training for high fidelity natural image synthesis,2018 CSCD被引 9 次    
12.  Lin C H. COCO-GAN: Generation by parts via conditional coordinating. 2019 IEEE/CVF International Conference on Computer Vision (ICCV),2019:4512-4521 CSCD被引 1 次    
13.  Karras T. A style-based generator architecture for generative adversarial networks. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),2019:4401-4410 CSCD被引 1 次    
14.  Karras T. Analyzing and improving the image quality of StyleGAN. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),2020:8110-8119 CSCD被引 1 次    
15.  何沛松. 面向GAN生成图像的被动取证及反取证技术综述. 中国图象图形学报,2022,27(1):88-110 CSCD被引 6 次    
16.  郑远攀. 深度学习在图像识别中的应用研究综述. 计算机工程与应用,2019,55(12):20-36 CSCD被引 103 次    
17.  Mo H X. Fake faces identification via convolutional neural network. Proceedings of the 6th ACM Workshop on Information Hiding and Multimedia Security,2018:43-47 CSCD被引 5 次    
18.  Dang L. Deep learning based computer generated face identification using convolutional neural network. Applied Sciences,2018,8(12):2610 CSCD被引 3 次    
19.  Zhang X. Detecting and simulating artifacts in GAN fake images. 2019 IEEE International Workshop on Information Forensics and Security (WIFS),2019:1-6 CSCD被引 2 次    
20.  Yu N. Attributing fake images to GANs: Learning and analyzing GAN fingerprints. 2019 IEEE/CVF International Conference on Computer Vision (ICCV),2019:7556-7566 CSCD被引 1 次    
引证文献 1 篇

1 谢天圻 GAN模型生成图像检测方法综述 计算机工程与应用,2024,60(22):74-86
CSCD被引 0 次

显示所有1篇文献

论文科学数据集
PlumX Metrics
相关文献

 作者相关
 关键词相关
 参考文献相关

版权所有 ©2008 中国科学院文献情报中心 制作维护：中国科学院文献情报中心
地址：北京中关村北四环西路33号 邮政编码：100190 联系电话：(010)82627496 E-mail:cscd@mail.las.ac.cn 京ICP备05002861号-4 | 京公网安备11010802043238号