FakeLocator: Robust Localization of GAN-Based Face Manipulations

被引:47
作者
Huang, Yihao [1 ]
Juefei-Xu, Felix [2 ]
Guo, Qing [3 ]
Liu, Yang [3 ,4 ]
Pu, Geguang [1 ,5 ]
机构
[1] East China Normal Univ, Sch Software Engn, Shanghai 200050, Peoples R China
[2] Alibaba Grp, Sunnyvale, CA 94085 USA
[3] Nanyang Technol Univ, Sch Comp Sci & Engn, Singapore 639798, Singapore
[4] Zhejiang Sci Tech Univ, Sch Comp Sci & Technol, Hangzhou 314423, Peoples R China
[5] Shanghai Ind Control Safety Innovat Technol Co Lt, Shanghai 200331, Peoples R China
基金
新加坡国家研究基金会;
关键词
Faces; Location awareness; Videos; Information integrity; Robustness; Forensics; Image resolution; DeepFake; face manipulation; DeepFake detection and localization; NETWORKS;
D O I
10.1109/TIFS.2022.3141262
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Full face synthesis and partial face manipulation by virtue of the generative adversarial networks (GANs) and its variants have raised wide public concerns. In the multi-media forensics area, detecting and ultimately locating the image forgery has become an imperative task. In this work, we investigate the architecture of existing GAN-based face manipulation methods and observe that the imperfection of upsampling methods therewithin could be served as an important asset for GAN-synthesized fake image detection and forgery localization. Based on this basic observation, we have proposed a novel approach, termed FakeLocator, to obtain high localization accuracy, at full resolution, on manipulated facial images. To the best of our knowledge, this is the very first attempt to solve the GAN-based fake localization problem with a gray-scale fakeness map that preserves more information of fake regions. To improve the universality of FakeLocator across multifarious facial attributes, we introduce an attention mechanism to guide the training of the model. To improve the universality of FakeLocator across different DeepFake methods, we propose partial data augmentation and single sample clustering on the training images. Experimental results on popular FaceForensics++, DFFD datasets and seven different state-of-the-art GAN-based face generation methods have shown the effectiveness of our method. Compared with the baselines, our method performs better on various metrics. Moreover, the proposed method is robust against various real-world facial image degradations such as JPEG compression, low-resolution, noise, and blur.
引用
收藏
页码:2657 / 2672
页数:16
相关论文
共 54 条
[1]  
Afchar D, 2018, IEEE INT WORKS INFOR
[2]  
Agarwal S., 2019, P IEEE CVF C COMP VI, P38, DOI [10.4108/eai.18-7-2019, DOI 10.4108/EAI.18-7-2019]
[3]  
[Anonymous], INT C LEARNING REPRE
[4]  
Center on the Developing Child at Harvard University, 2017, Faceswap
[5]   What Makes Fake Images Detectable? Understanding Properties that Generalize [J].
Chai, Lucy ;
Bau, David ;
Lim, Ser-Nam ;
Isola, Phillip .
COMPUTER VISION - ECCV 2020, PT XXVI, 2020, 12371 :103-120
[6]   StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation [J].
Choi, Yunjey ;
Choi, Minje ;
Kim, Munyoung ;
Ha, Jung-Woo ;
Kim, Sunghun ;
Choo, Jaegul .
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, :8789-8797
[7]   Xception: Deep Learning with Depthwise Separable Convolutions [J].
Chollet, Francois .
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :1800-1807
[8]   On the Detection of Digital Face Manipulation [J].
Dang, Hao ;
Liu, Feng ;
Stehouwer, Joel ;
Liu, Xiaoming ;
Jain, Anil K. .
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2020, :5780-5789
[9]   Deep Learning Based Computer Generated Face Identification Using Convolutional Neural Network [J].
Dang, L. Minh ;
Hassan, Syed Ibrahim ;
Im, Suhyeon ;
Lee, Jaecheol ;
Lee, Sujin ;
Moon, Hyeonjoon .
APPLIED SCIENCES-BASEL, 2018, 8 (12)
[10]  
Frank J, 2020, PR MACH LEARN RES, V119