Filters








270 Hits in 5.8 sec

Detecting GAN generated Fake Images using Co-occurrence Matrices [article]

Lakshmanan Nataraj, Tajuddin Manhar Mohammed, Shivkumar Chandrasekaran, Arjuna Flenner, Jawadul H. Bappy, Amit K. Roy-Chowdhury, B. S. Manjunath
2019 arXiv   pre-print
In this paper, we propose a novel approach to detect GAN generated fake images using a combination of co-occurrence matrices and deep learning.  ...  We extract co-occurrence matrices on three color channels in the pixel domain and train a model using a deep convolutional neural network (CNN) framework.  ...  Conclusions In this paper, we proposed a novel method to detect GAN generated fake images using a combination of pixel co-occurrence matrices and deep learning.  ... 
arXiv:1903.06836v2 fatcat:ynjrs77o7ramlav3bhuntcgkfa

Detection, Attribution and Localization of GAN Generated Images [article]

Michael Goebel, Lakshmanan Nataraj, Tejaswi Nanjundaswamy, Tajuddin Manhar Mohammed, Shivkumar Chandrasekaran, B.S. Manjunath
2020 arXiv   pre-print
For every image, co-occurrence matrices are computed on neighborhood pixels of RGB channels in different directions (horizontal, vertical and diagonal).  ...  GANs are used in a wide range of tasks, from modifying small attributes of an image (StarGAN [14]), transferring attributes between image pairs (CycleGAN [91]), as well as generating entirely new images  ...  For detection, we consider a two class framework -real and GAN, where a network is trained on co-occurrence matrices computed on the whole image to detect if an image is real or GAN generated.  ... 
arXiv:2007.10466v1 fatcat:dt3uho2im5exhcvhfy2qdasrxy

CNN Detection of GAN-Generated Face Images based on Cross-Band Co-occurrences Analysis [article]

Mauro Barni, Kassem Kallas, Ehsan Nowroozi, Benedetta Tondi
2020 arXiv   pre-print
Specifically, we use cross-band co-occurrence matrices, in addition to spatial co-occurrence matrices, as input to a CNN model, which is trained to distinguish between real and synthetic faces.  ...  Last-generation GAN models allow to generate synthetic images which are visually indistinguishable from natural ones, raising the need to develop tools to distinguish fake and natural images thus contributing  ...  to distinguish between real and fake (GAN-generated) images.  ... 
arXiv:2007.12909v2 fatcat:wi4uhgnjgnfndlsnygzoq6lgju

VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

Anselmo Ferreira, Ehsan Nowroozi, Mauro Barni
2021 Journal of Imaging  
Additionally, printing and scanning can be used to hide the traces of image manipulation or the synthetic nature of images, since the artifacts commonly found in manipulated and synthetic images are gone  ...  First of all, printed documents are often associated with criminal activities, such as terrorist plans, child pornography, and even fake packages.  ...  Co-occurrence matrices proposed in [39] to discriminate GAN-generated images from natural ones and their behavior in digital and printed and scanned images: (top) GAN image cooccurrence matrices in the  ... 
doi:10.3390/jimaging7030050 pmid:34460706 pmcid:PMC8321306 fatcat:5hch27by45frrmevvtquczmw2i

Adversarial Attacks on Co-Occurrence Features for GAN Detection [article]

Michael Goebel, B. S. Manjunath
2020 arXiv   pre-print
With this rise in ability to generate fake images comes demand to detect them. While numerous methods have been developed for this task, the majority of them remain vulnerable to adversarial attacks.  ...  In this paper, develop two novel adversarial attacks on co-occurrence based GAN detectors. These are the first attacks to be presented against such a detector.  ...  Generalizes the discrete co-occurrence matrices used for GAN detection to a differentiable function, and Figure 1 . Outline of this paper's primary contribution, the graybox co-occurrence attack.  ... 
arXiv:2009.07456v1 fatcat:uencneivpvcfvdwl2bo55enlu4

VIPPrint: A Large Scale Dataset of Printed and Scanned Images for Synthetic Face Images Detection and Source Linking [article]

Anselmo Ferreira, Ehsan Nowroozi, Mauro Barni
2021 arXiv   pre-print
Additionally, printing and scanning can be used to hide the traces of image manipulation or the synthetic nature of images, since the artifacts commonly found in manipulated and synthetic images are gone  ...  First of all, printed documents are often associated with criminal activities, such as terrorist plans, child pornography pictures, and even fake packages.  ...  The experiments we have run guide us to a bunch of future works.  ... 
arXiv:2102.06792v1 fatcat:dhp6o6355re4zikz6y42odv3pm

One-Shot GAN Generated Fake Face Detection [article]

Hadi Mansourifar, Weidong Shi
2020 arXiv   pre-print
In this paper, we propose a universal One-Shot GAN generated fake face detection method which can be used in significantly different areas of anomaly detection.  ...  We prove that, the proposed method can outperform previous methods based on our experiments on Style-GAN generated fake faces.  ...  Co-occurrence matrices are computed on the color channels of an image and then trained a deep convolutional neural network to distinguish GAN generated fake images from real ones. Hsu et al.  ... 
arXiv:2003.12244v1 fatcat:z5lz7ldswfec5gbxteftcmyzza

Detection of Deep Network Generated Images Using Disparities in Color Components [article]

Haodong Li and Bin Li and Shunquan Tan and Jiwu Huang
2019 arXiv   pre-print
In this paper, we address the problem of detecting deep network generated (DNG) images by analyzing the disparities in color components between real scene images and DNG images.  ...  Based on these observations, we propose a feature set to capture color image statistics for detecting the DNG images.  ...  The truncated image residuals are then used to compute the co-occurrence matrices. 4) Extracting co-occurrence features: In total, we have five co-occurrence matrices, which are calculated fromR RGB  ... 
arXiv:1808.07276v2 fatcat:2susnsislnfnzf43sonfxp6c5e

DA-FDFtNet: Dual Attention Fake Detection Fine-tuning Network to Detect Various AI-Generated Fake Images [article]

Young Oh Bang, Simon S. Woo
2021 arXiv   pre-print
However, detecting realistic fake images generated by the latest AI technology is challenging due to the reasons mentioned above.  ...  Due to the advancement of Generative Adversarial Networks (GAN), Autoencoders, and other AI technologies, it has been much easier to create fake images such as "Deepfakes".  ...  to learn important co-occurrence matrices essential fea- tures.  ... 
arXiv:2112.12001v1 fatcat:b7wmsotrqndqhd5e7j4cigiray

Face Generation using Deep Convolutional Generative Adversarial Neural Network

Devaki P
2020 Bioscience Biotechnology Research Communications  
GAN typically work with image dataset but they are difficult to train. This paper explores the potential of GAN to generate realistic images.  ...  CNNs are used for feature detections by looking at the image and try to check if certain features are present in the image and then it classifies the image accordingly.  ...  (Nataraj et al,.2019) Detecting Fake images using Cooccurrence matrices is GAN based technique to identify fake images such as Deep fakes, Image-to-Image transitions.  ... 
doi:10.21786/bbrc/13.11/5 fatcat:ytjwwxepyjevfaw7orrp7xbcme

Are GAN generated images easy to detect? A critical analysis of the state-of-the-art [article]

Diego Gragnaniello, Davide Cozzolino, Francesco Marra, Giovanni Poggi, Luisa Verdoliva
2021 arXiv   pre-print
In this work, we analyze the state-of-the-art methods for the detection of synthetic images, highlighting the key ingredients of the most successful approaches, and comparing their performance over existing  ...  generative architectures.  ...  However, this technology can also be used for malicious purposes, for example to generate fake profiles on social network or to generate fake news.  ... 
arXiv:2104.02617v1 fatcat:fuocaj263fh5xoo2ui4tsisjmm

Detection of GAN-Synthesized Image Based on Discrete Wavelet Transform

Guihua Tang, Lei Sun, Xiuqing Mao, Song Guo, Hongmeng Zhang, Xiaoqin Wang, Beijing Chen
2021 Security and Communication Networks  
In this study, a method named fake images discriminator (FID) is proposed, which detects that GAN-synthesized fake images use the strong spectral correlation in the imaging process of natural color images  ...  Recently, generative adversarial networks (GANs) and its variants have shown impressive ability in image synthesis.  ...  [3] built a pixel-level image detection model based on the deep neural network (DNN) and detected GAN-synthesized images by extract co-occurrence matrices on three color channels in the pixel domain  ... 
doi:10.1155/2021/5511435 fatcat:jsgwhbmzsbafnlf2dyrcqssjcy

Challenges and Solutions in DeepFakes [article]

Jatin Sharma, Sahil Sharma
2021 arXiv   pre-print
1 million fake faces generated by style GAN.  ...  It helps to create fake images and videos that human cannot distinguish them from the real ones and are recent off-shelf manipulation technique that allows swapping two identities in a single video.  ...  [44] CNN Deep fea- tures + co- occurrences matrices Cycle GAN Work with static image only Yu et al. [45] CNN Deep fea- tures CelebA Poor performance on post- processing operations.  ... 
arXiv:2109.05397v2 fatcat:lhs5awtsy5ghxctxtcknd4lm7y

Perception Matters: Exploring Imperceptible and Transferable Anti-forensics for GAN-generated Fake Face Imagery Detection [article]

Yongwei Wang, Xin Ding, Li Ding, Rabab Ward, Z. Jane Wang
2020 arXiv   pre-print
Recently, generative adversarial networks (GANs) can generate photo-realistic fake facial images which are perceptually indistinguishable from real face photos, promoting research on fake face detection  ...  Though fake face forensics can achieve high detection accuracy, their anti-forensic counterparts are less investigated.  ...  Then ensemble steganalysis classifiers were employed using features extracted from a third order co-occurrence matrix.  ... 
arXiv:2010.15886v1 fatcat:emsfw5a5wvfrfgmtls4i5nexfe

Detecting CNN-Generated Facial Images in Real-World Scenarios [article]

Nils Hulzebosch, Sarah Ibrahimi, Marcel Worring
2020 arXiv   pre-print
Furthermore, we examine the usefulness of commonly used image pre-processing methods.  ...  Artificial, CNN-generated images are now of such high quality that humans have trouble distinguishing them from real images.  ...  Lastly, several works use co-occurrence matrices to focus on ir-regularities in pixel-patterns, for example in steganalysis [12, 14, 38, 45, 46] and detection of forged images [11, 12] .  ... 
arXiv:2005.05632v1 fatcat:dxjzgl5wknagrpx643c3eh2lly
« Previous Showing results 1 — 15 out of 270 results