Filters








205 Hits in 4.4 sec

Training Strategies and Data Augmentations in CNN-based DeepFake Video Detection [article]

Luca Bondi, Edoardo Daniele Cannas, Paolo Bestagini, Stefano Tubaro
2020 arXiv   pre-print
In this paper we analyze how different training strategies and data augmentation techniques affect CNN-based deepfake detectors when training and testing on the same dataset or across different datasets  ...  The fast and continuous growth in number and quality of deepfake videos calls for the development of reliable detection systems capable of automatically warning users on social media and on the Internet  ...  In light of the results in terms of single augmentation, we build a data augmentation pipeline based on HF, BC, HSV, and JPEG, and re-train the CNN with both BCE and triplet loss.  ... 
arXiv:2011.07792v1 fatcat:obhbfdtngzbotkuiufe5pbnvpi

Training Strategies and Data Augmentations in CNN-based DeepFake Video Detection

Luca Bondi, Edoardo Daniele Cannas, Paolo Bestagini, Stefano Tubaro
2020 2020 IEEE International Workshop on Information Forensics and Security (WIFS)  
In this paper we analyze how different training strategies and data augmentation techniques affect CNN-based deepfake detectors when training and testing on the same dataset or across different datasets  ...  The fast and continuous growth in number and quality of deepfake videos calls for the development of reliable detection systems capable of automatically warning users on social media and on the Internet  ...  In light of the results in terms of single augmentation, we build a data augmentation pipeline based on HF, BC, HSV, and JPEG, and re-train the CNN with both BCE and triplet loss.  ... 
doi:10.1109/wifs49906.2020.9360901 fatcat:rw3hdkukxvhzzeos53oo35vu5y

Comparison of Deepfake Detection Techniques through Deep Learning

Maryam Taeb, Hongmei Chi
2022 Journal of Cybersecurity and Privacy  
In this work, we compare the most common, state-of-the-art face-detection classifiers such as Custom CNN, VGG19, and DenseNet-121 using an augmented real and fake face-detection dataset.  ...  in video and digital content).  ...  Their insight and expertise and exacting attention to detail has greatly assisted this research Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/jcp2010007 fatcat:t7blxnyr65a73nlot4lukg32g4

Exposing DeepFake Videos By Detecting Face Warping Artifacts [article]

Yuezun Li, Siwei Lyu
2019 arXiv   pre-print
Since training a DeepFake model to generate negative examples is time-consuming and resource-demanding, our method saves a plenty of time and resources in training data collection; (2) Since such artifacts  ...  Such transforms leave distinctive artifacts in the resulting DeepFake videos, and we show that they can be effectively captured by convolutional neural networks (CNNs).  ...  The views, opinions, and findings expressed are those of the authors and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government.  ... 
arXiv:1811.00656v3 fatcat:l2in3ilzprfzpg5iqvbjbt6s2e

A New Approach to Improve Learning-based Deepfake Detection in Realistic Conditions [article]

Yuhang Lu, Touradj Ebrahimi
2022 arXiv   pre-print
This paper proposes a more effective data augmentation scheme based on real-world image degradation process.  ...  This novel technique is deployed for deepfake detection tasks and has been evaluated by a more realistic assessment framework.  ...  ILLUSTRATIVE EXAMPLE FOR DEEPFAKE DETECTION The usage of the proposed augmentation strategy is illustrated in the context of deepfake detection , which has become a hot topic in digital forensics. 1) Detection  ... 
arXiv:2203.11807v1 fatcat:koy3tzpkjredhduvsn5mt6go2q

FDFtNet: Facing Off Fake Images using Fake Detection Fine-tuning Network [article]

Hyeonseong Jeon, Youngoh Bang, Simon S. Woo
2020 arXiv   pre-print
This module is added to the pre-trained model and fine-tuned on a few data to search for new sets of feature space to detect fake images.  ...  Creating fake images and videos such as "Deepfake" has become much easier these days due to the advancement in Generative Adversarial Networks (GANs).  ...  The training strategies for Face2Face are very similar to those of the Deepfake dataset. Data augmentation is also applied. M , N , α, and β are set to 3, 4, 3, and 10, respectively.  ... 
arXiv:2001.01265v2 fatcat:6bfxfynfcvfllnuylwyyxhtbdy

Status, Challenges, and Future Views of DeepFake Techniques and Datasets

Ahmed S Abdulreda, Ahmed J Obaid
2022 Mathematical Statistician and Engineering Applications  
methods and methods in counterfeiting and anti-counterfeiting detection methods, we will address in this paper one of the methods that help in detecting image manipulation specifically in a modified manner  ...  on the GAN algorithm and we will call it ADVGAN and it includes the process of selecting and configuring a specific database Several initial operations and training by learning the machine and relying  ...  Introduction This paper discusses the strategies that deal with deepfaking and its types, the data used in manipulation, and the methods for detecting counterfeiting.  ... 
doi:10.17762/msea.v71i2.81 fatcat:bx4zhxerjfhfzdjsy22r6tsto4

ID-Reveal: Identity-aware DeepFake Video Detection [article]

Davide Cozzolino and Andreas Rössler and Justus Thies and Matthias Nießner and Luisa Verdoliva
2021 arXiv   pre-print
A major challenge in DeepFake forgery detection is that state-of-the-art algorithms are mostly trained to detect a specific fake method.  ...  The advantage is that we do not need any training data of fakes, but only train on real videos.  ...  In addition, this material is based on research sponsored by the Defense Advanced Research Projects Agency (DARPA) and the Air Force Research Laboratory (AFRL) under agreement number FA8750-20-2-1004.  ... 
arXiv:2012.02512v3 fatcat:j7do67ktenbstoye66xucybfsy

A Convolutional LSTM based Residual Network for Deepfake Video Detection [article]

Shahroz Tariq, Sangyup Lee, Simon S. Woo
2020 arXiv   pre-print
In recent years, deep learning-based video manipulation methods have become widely accessible to masses.  ...  Several deep learning-based detection methods have been developed to identify these deepfakes.  ...  We noticed that most of the deepfake detection methods [10, 35] randomly extract frames (images) from videos for training and testing, hence a single frame-based detection method.  ... 
arXiv:2009.07480v1 fatcat:5jcj3dd7jnakrkhv3glwzkjo4u

BZNet: Unsupervised Multi-scale Branch Zooming Network for Detecting Low-quality Deepfake Videos

Sangyup Lee, Jaeju An, Simon S. Woo
2022 Proceedings of the ACM Web Conference 2022  
Extensive experiments on the FaceForensics++ LQ and GAN-generated datasets demonstrate that our BZNet architecture improves the detection accuracy of existing CNN-based classifiers by 4.21% on average.  ...  We train our BZNet only using highly compressed LQ images and experiment under a realistic setting, where HQ training data are not readily accessible.  ...  To diversify the training dataset, we employed image data augmentation techniques: 1) hue, saturation, brightness, and contrast (-30% to 30%), 2) horizontal flip (50% probability), and 3) rotation (-20  ... 
doi:10.1145/3485447.3512245 fatcat:pprtqq4z2zbnbecpzvfgflwxau

Detecting Deepfake Images Using Deep Learning Techniques and Explainable AI Methods

Wahidul Hasan Abir, Faria Rahman Khanam, Kazi Nabiul Alam, Myriam Hadjouni, Hela Elmannai, Sami Bourouis, Rajesh Dey, Mohammad Monirujjaman Khan
2023 Intelligent Automation and Soft Computing  
Deepfake content is created with the help of artificial intelligence and machine learning to replace one person's likeness with another person in pictures or recorded videos.  ...  Although visual media manipulations are not new, the introduction of deepfakes has marked a breakthrough in creating fake media and information.  ...  Lastly, horizontal flips were added in the data augmentation phase to make the model more robust for unseen test images. The data augmentation was applied to the train, test, and validation sets.  ... 
doi:10.32604/iasc.2023.029653 fatcat:cfjydbkat5bjncom66fiqa56em

A Novel Framework for Assessment of Learning-based Detectors in Realistic Conditions with Application to Deepfake Detection [article]

Yuhang Lu, Ruizhi Luo, Touradj Ebrahimi
2022 arXiv   pre-print
Inspired by the assessment results, a data augmentation strategy based on natural image degradation process is designed, which significantly improves the generalization ability of two deepfake detectors  ...  This paper proposes a rigorous framework to assess performance of learning-based detectors in more realistic situations. An illustrative example is shown under deepfake detection context.  ...  In the end, we have designed a data augmentation strategy based on realistic image degradation modeling process, which significantly improves the generalization ability of a deepfake detector at marginal  ... 
arXiv:2203.11797v1 fatcat:2mbwo2sslvb2dlswteha2msjgu

The eyes know it: FakeET – An Eye-tracking Database to Understand Deepfake Perception [article]

Parul Gupta, Komal Chugh, Abhinav Dhall, Ramanathan Subramanian
2020 arXiv   pre-print
The compiled data confirms (a) distinct eye movement characteristics for real vs fake videos; (b) utility of the eye-track saliency maps for spatial forgery localization and detection, and (c) Error Related  ...  Negativity (ERN) triggers in the EEG responses, and the ability of the raw EEG signal to distinguish between real and fake videos.  ...  To empirically validate these observations, we trained a CNN to predict if a video is real or fake. The base of the CNN in our study is a standard 3D ResNet [7] .  ... 
arXiv:2006.06961v2 fatcat:5ppqezyczjaojhanmn25il3lyy

Robust Deepfake On Unrestricted Media: Generation And Detection [article]

Trung-Nghia Le and Huy H Nguyen and Junichi Yamagishi and Isao Echizen
2022 arXiv   pre-print
It also discusses possible ways to improve the robustness of deepfake detection for a wide variety of media (e.g., in-the-wild images and videos).  ...  This chapter explores the evolution of and challenges in deepfake generation and detection.  ...  Acknowledgments This research was partly supported by JSPS KAKENHI Grants (JP16H06302, JP18H04120, JP21H04907, JP20K23355, JP21K18023) and JST CREST Grants (JPMJCR20D3, JP-MJCR18A6), Japan.  ... 
arXiv:2202.06228v1 fatcat:a37q2lf7w5bcbekk5esmbx2goe

DeepFake Detection for Human Face Images and Videos: A Survey

Asad Malik, Minoru Kuribayashi, Sani M. Abdullahi, Ahmad Neyaz Khan
2022 IEEE Access  
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.  ...  We hope that the knowledge encompassed in this survey will accelerate the use of deep learning in face image and video DeepFake detection methods.  ...  [111] , [112] proposed a method to identify human facial expressions using data augmentation and fine-tuning the CNN model.  ... 
doi:10.1109/access.2022.3151186 fatcat:imz6hdtofrbxfcfi6kput2mffi
« Previous Showing results 1 — 15 out of 205 results