Filters








234 Hits in 9.3 sec

Detecting GAN-generated Imagery using Color Cues [article]

Scott McCloskey, Michael Albright
2018 arXiv   pre-print
We further show that these two cues can be used to distinguish GAN-generated imagery from camera imagery, demonstrating effective discrimination between GAN imagery and real camera images used to train  ...  Of particular interest, given its recent successes, is the detection of imagery produced by Generative Adversarial Networks (GANs), e.g. 'deepfakes'.  ...  We investigate the effectiveness of these two cues in detecting two types of GAN imagery: one being imagery wholly generated by a GAN and the other where GAN-generated faces replace real faces in a larger  ... 
arXiv:1812.08247v1 fatcat:a5zsqgqar5awxfqqvg74xqpdnu

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

Nils Hulzebosch, Sarah Ibrahimi, Marcel Worring
2020 arXiv   pre-print
detection methods using the proposed framework.  ...  Our results suggest that CNN-based detection methods are not yet robust enough to be used in real-world scenarios.  ...  GANs [16] have recently emerged as the state-of-the-art in generating realistic imagery, in terms of image resolution and visual quality.  ... 
arXiv:2005.05632v1 fatcat:dxjzgl5wknagrpx643c3eh2lly

Manipulation Detection in Satellite Images Using Deep Belief Networks [article]

János Horváth, Daniel Mas Montserrat, Hanxiang Hao, Edward J. Delp
2020 arXiv   pre-print
Image manipulation tools including both manual editing tools and automated techniques can be easily used to tamper and modify satellite imagery.  ...  In this paper, we present a one-class detection method based on deep belief networks (DBN) for splicing detection and localization without using any prior knowledge of the manipulations.  ...  Some of these techniques include detecting tampering by finding double-JPEG compression artifacts [31] , using neural networks with domain adaptation [32] or using saturation cues [33] .  ... 
arXiv:2004.12441v1 fatcat:efjllje4frgnnham2qs7wdufwm

Manipulation Detection in Satellite Images Using Deep Belief Networks

Janos Horvath, Daniel Mas Montserrat, Hanxiang Hao, Edward J. Delp
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
Image manipulation tools including both manual editing tools and automated techniques can be easily used to tamper and modify satellite imagery.  ...  In this paper, we present a one-class detection method based on deep belief networks (DBN) for splicing detection and localization without using any prior knowledge of the manipulations.  ...  Some of these techniques include detecting tampering by finding double-JPEG compression artifacts [31] , using neural networks with domain adaptation [32] or using saturation cues [33] .  ... 
doi:10.1109/cvprw50498.2020.00340 dblp:conf/cvpr/HorvathMHD20 fatcat:m2ku44o7j5gohpde6culchb2tq

Manipulation Detection in Satellite Images Using Vision Transformer [article]

János Horváth, Sriram Baireddy, Hanxiang Hao, Daniel Mas Montserrat, Edward J. Delp
2021 arXiv   pre-print
Overhead imagery is used by numerous industries including agriculture, forestry, natural disaster analysis, and meteorology.  ...  In this paper we propose an unsupervised technique that uses a Vision Transformer to detect spliced areas within satellite images.  ...  the Radon transform of resampled features and a deep learning classifier [4] , and using saturation cues [35] .  ... 
arXiv:2105.06373v1 fatcat:gdz4x2idkfanzcpi5te563k66e

Detecting CNN-Generated Facial Images in Real-World Scenarios

Nils Hulzebosch, Sarah Ibrahimi, Marcel Worring
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
detection methods using the proposed framework.  ...  Our results suggest that CNN-based detection methods are not yet robust enough to be used in real-world scenarios.  ...  GANs [16] have recently emerged as the state-of-the-art in generating realistic imagery, in terms of image resolution and visual quality.  ... 
doi:10.1109/cvprw50498.2020.00329 dblp:conf/cvpr/HulzeboschIW20 fatcat:ibi7ctk6rzev5djd2wrzirq5b4

Machine Vision for Improved Human-Robot Cooperation in Adverse Underwater Conditions [article]

Md Jahidul Islam
2021 arXiv   pre-print
In addition to advancing the state-of-the-art, the proposed methodologies and systems take us one step closer toward bridging the gap between theory and practice for improved human-robot cooperation in  ...  from over-saturation, while LS-GAN generally fails to rectify the greenish hue in images.UGAN-P, Pix2Pix, and Uw-HL perform reasonably well and their enhanced images are comparable to that of FUnIE-GAN  ...  First, FUnIE-GAN is not very effective for enhancing severely degraded texture-less images. The generated images in such cases are often over-saturated by noise amplification.  ... 
arXiv:1911.07623v2 fatcat:tb7voaqrlbgtfgzvdwvjfr64vu

Object Tracking Based on Satellite Videos: A Literature Review

Zhaoxiang Zhang, Chenghang Wang, Jianing Song, Yuelei Xu
2022 Remote Sensing  
Finally, a revised multi-level dataset based on wpafb videos is generated and quantitatively evaluated for future development in the satellite video-based tracking area.  ...  Acknowledgments: We thank the anonymous reviewers and editors for their constructive comments and suggestions, which helped us to improve the manuscript.  ...  GAN-Based Tracking Methods Models such as those above can be categorized as discriminative models as they use conditional probability to predict the unseen data, while other methods employ generative models  ... 
doi:10.3390/rs14153674 fatcat:fhpk7dx6iba55msd3o2kaxppa4

2D GANs Meet Unsupervised Single-view 3D Reconstruction [article]

Feng Liu, Xiaoming Liu
2022 arXiv   pre-print
In light of this, we propose a novel image-conditioned neural implicit field, which can leverage 2D supervisions from GAN-generated multi-view images and perform the single-view reconstruction of generic  ...  Recent research has shown that controllable image generation based on pre-trained GANs can benefit a wide range of computer vision tasks. However, less attention has been devoted to 3D vision tasks.  ...  To first generate multi-view imagery by a pre-trained GAN, e.g., StyleGAN, we carefully study the latent space of StyleGAN and devise a simple but effective technique, which generates plausible images  ... 
arXiv:2207.10183v1 fatcat:mqxluaenc5ekrg3wbusbnhumsa

Generating Triples With Adversarial Networks for Scene Graph Construction

Matthew Klawonn, Eric Heim
2018 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Despite this, there have been only a few recent efforts to generate scene graphs from imagery.  ...  We take the approach of first generating small subgraphs, each describing a single statement about a scene from a specific region of the input image chosen using an attention mechanism.  ...  The visual cues necessary to detect relations and those necessary to detect attributes vary significantly, increasing the burden on any architecture attempting to capture both at the same time.  ... 
doi:10.1609/aaai.v32i1.12321 fatcat:2bpofvgtinbk5jxluazz6esequ

Simultaneous Enhancement and Super-Resolution of Underwater Imagery for Improved Visual Perception [article]

Md Jahidul Islam, Peigen Luo, Junaed Sattar
2020 arXiv   pre-print
We present Deep SESR, a residual-in-residual network-based generative model that can learn to restore perceptual image qualities at 2x, 3x, or 4x higher spatial resolution.  ...  We also validate its generalization performance on several test cases that include underwater images with diverse spectral and spatial degradation levels, and also terrestrial images with unseen natural  ...  We are also grateful to the Bellairs Research Institute 2 of Barbados for providing us with the facilities for field experiments. Additionally, we thank Nvidia TM for donating two GPUs for our work.  ... 
arXiv:2002.01155v1 fatcat:xqurixesvvfj7gpebhkx2x3tgm

Generating Triples with Adversarial Networks for Scene Graph Construction [article]

Matthew Klawonn, Eric Heim
2018 arXiv   pre-print
Despite this, there have been only a few recent efforts to generate scene graphs from imagery.  ...  We take the approach of first generating small subgraphs, each describing a single statement about a scene from a specific region of the input image chosen using an attention mechanism.  ...  The visual cues necessary to detect relations and those necessary to detect attributes vary significantly, increasing the burden on any architecture attempting to capture both at the same time.  ... 
arXiv:1802.02598v1 fatcat:3i5s3x2zzjhbtn2neqowam5n5y

Survey on Videos Data Augmentation for Deep Learning Models

Nino Cauli, Diego Reforgiato Recupero
2022 Future Internet  
More advanced solutions are Domain Randomization methods or the use of simulation to artificially generate the missing data.  ...  Data augmentation can alleviate the problem, generating new data from a smaller initial dataset.  ...  These strategies use Generative adversarial networks (GANs) to generate the new images [9] .  ... 
doi:10.3390/fi14030093 fatcat:jdruypthqfddho55hzwrcypxae

DeepFake Detection for Human Face Images and Videos: A Survey

Asad Malik, Minoru Kuribayashi, Sani M. Abdullahi, Ahmad Neyaz Khan
2022 IEEE Access  
Additionally, the issue of how DeepFake detection aims to generate a generalized DeepFake detection model will be analyzed.  ...  In this survey, we will summarize the DeepFake detection methods in face images and videos on the basis of their results, performance, methodology used and detection type.  ...  GENERATOR C. GANs background GANs are a revolutionary tool used for teaching generative models to generate realistic examples from a data distribution [2] .  ... 
doi:10.1109/access.2022.3151186 fatcat:imz6hdtofrbxfcfi6kput2mffi

How Do the Hearts of Deep Fakes Beat? Deep Fake Source Detection via Interpreting Residuals with Biological Signals [article]

Umur Aybars Ciftci and Ilke Demir and Lijun Yin
2020 arXiv   pre-print
Following these generation techniques, some detection approaches have also been proved useful due to their high classification accuracy.  ...  Some pure deep learning based approaches try to classify deep fakes using CNNs where they actually learn the residuals of the generator.  ...  Deep Fake Detectors For fake image detection from the face generation category, several typical signatures have been identified including saturation cues [46] , frequencies of generated images for fingerprints  ... 
arXiv:2008.11363v1 fatcat:h7tt4rw4uzfrbeoei5faqfxbey
« Previous Showing results 1 — 15 out of 234 results