Deep learning for deepfakes creation and detection: A survey

@article{Nguyen2019DeepLF,
  title={Deep learning for deepfakes creation and detection: A survey},
  author={Thanh Thi Nguyen and Quoc Viet Hung Nguyen and Dung Nguyen and Duc Thanh Nguyen and Thien Huynh-The and Saeid Nahavandi and Thanh Tam Nguyen and Quoc-Viet Pham and Cu Nguyen},
  journal={Comput. Vis. Image Underst.},
  year={2019},
  volume={223},
  pages={103525}
}
[PDF] Semantic Reader
114 Citations
Highly Influential Citations
7
Background Citations
57
Methods Citations
24

Figures and Tables from this paper

114 Citations

Towards Generalizable Deepfake Detection with Locality-aware AutoEncoder

Locality-Aware AutoEncoder (LAE) is proposed to bridge the generalization gap in deepfakes by using a pixel-wise mask to regularize local interpretation of LAE to enforce the model to learn intrinsic representation from the forgery region, instead of capturing artifacts in the training set and learning superficial correlations to perform detection.

A Study on Combating Emerging Threat of Deepfake Weaponization

  • R. KataryaAnushka Lal
  • Computer Science
    2020 Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
  • 2020
SSTNet is presented as the best model to date, that uses spatial, temporal, and steganalysis for detection of deepfakes and the threat posed by document and signature forgery is highlighted.

Two-Branch Recurrent Network for Isolating Deepfakes in Videos

A method for deepfake detection based on a two-branch network structure that isolates digitally manipulated faces by learning to amplify artifacts while suppressing the high-level face content, and derives a novel cost function that compresses the variability of natural faces and pushes away the unrealistic facial samples in the feature space.

Approaches for Fake Content Detection: Strengths and Weaknesses to Adversarial Attacks

Examples of machine and deep learning approaches that can be used to detect different types of fake content are described and some future research challenges in this area are identified.

Integrating Deepfake Detection into Cybersecurity Curriculum

A set of hands-on labs to integrate them in the cybersecurity curriculum so that students can be educated to use detect software and identify Deepfakes, and the fundamental capabilities, challenges, and limitations of deep learning for detecting smart attacks are investigated.

Deepfake Detection using Spatiotemporal Convolutional Networks

This work created a benchmark of the performance of spatiotemporal convolutional methods using the Celeb-DF dataset and found that the methods outperformed state-of-the-art frame-based detection methods.

Deep Learning Methods for Credit Card Fraud Detection

Experimental results show great performance of the proposed deep learning methods against traditional machine learning models and imply that the proposed approaches can be implemented effectively for real-world credit card fraud detection systems.

Towards measuring fairness in AI: the Casual Conversations dataset

A novel dataset to help researchers evaluate their computer vision and audio models for accuracy across a diverse set of age, genders, apparent skin tones and ambient lighting conditions and a thorough analysis on the top five winners of the DeepFake Detection Challenge is provided.

Deep Reinforcement Learning for Cyber Security

  • T. NguyenV. Reddi
  • Computer Science
    IEEE transactions on neural networks and learning systems
  • 2021
This article presents a survey of DRL approaches developed for cyber security, including DRL-based security methods for cyber-physical systems, autonomous intrusion detection techniques, and multiagent D RL-based game theory simulations for defense strategies against cyberattacks.

Intermediate Layer Optimization for Inverse Problems using Deep Generative Models

This work proposes Intermediate Layer Optimization (ILO), a novel optimization algorithm for solving inverse problems with deep generative models that outperforms state-of-the-art methods introduced in StyleGAN-2 and PULSE for a wide range of inverse problems.
...

References

SHOWING 1-10 OF 182 REFERENCES

DeepFake Detection by Analyzing Convolutional Traces

This work focuses on the analysis of Deepfakes of human faces with the objective of creating a new detection method able to detect a forensics trace hidden in images: a sort of fingerprint left in the image generation process.

Interpretable Deepfake Detection via Dynamic Prototypes

This work proposes Dynamic Prototype Network (DPNet) - a simple, interpretable, yet effective solution that leverages dynamic representations (i.e., prototypes) to explain deepfake visual dynamics.

DeepFakes: a New Threat to Face Recognition? Assessment and Detection

This paper presents the first publicly available set of Deepfake videos generated from videos of VidTIMIT database, and demonstrates that GAN-generated Deep fake videos are challenging for both face recognition systems and existing detection methods.

A Deep Learning Approach to Universal Image Manipulation Detection Using a New Convolutional Layer

A universal forensic approach to performing manipulation detection using deep learning that can automatically learn how to detect multiple image manipulations without relying on pre-selected features or any preprocessing is proposed.

Effective and Fast DeepFake Detection Method Based on Haar Wavelet Transform

  • M. YounusT. Hasan
  • Computer Science
    2020 International Conference on Computer Science and Software Engineering (CSASE)
  • 2020
The proposed DeepFake detection method takes the advantage of the fact that current DeepFake generation algorithms cannot generate face images with varied resolutions, it is only able to generate new faces with a limited size and resolution, a further distortion and blur is needed to match and fit the fake face with the background and surrounding context in the source video.

Recurrent Convolutional Structures for Audio Spoof and Video Deepfake Detection

This work introduces simple, yet surprisingly efficient digital forensic methods for audio spoof and visual deepfake detection that combine convolutional latent representations with bidirectional recurrent structures and entropy-based cost functions.

Vulnerability assessment and detection of Deepfake videos

This paper presents the first publicly available set of Deepfake videos generated from videos of VidTIMIT database, and demonstrates that GAN-generated Deep fake videos are challenging for both face recognition systems and existing detection methods.

Exposing DeepFake Videos By Detecting Face Warping Artifacts

A new deep learning based method that can effectively distinguish AI-generated fake videos from real videos is described, which saves a plenty of time and resources in training data collection and is more robust compared to others.

Deep Fake Image Detection Based on Pairwise Learning

This paper proposes a deep learning-based approach for detecting the fake images by using the contrastive loss and demonstrated that the proposed method significantly outperformed other state-of-the-art fake image detectors.

Blind detection of glow-based facial forgery

A Convolutional Neural Network (CNN) based model, namely SCnet, is proposed to expose the Glow-based facial forgery and is specifically designed to automatically learn high-level forensics features from pre-processing results.
...