Media Forensics and Deepfake Systematic Survey

View on arXiv ← Back to list

Authors: Nadeem Jabbar CH, Aqib Saghir, Ayaz Ahmad Meer, Salman Ahmad Sahi, Bilal Hassan, Siddiqui Muhammad Yasir

Published: 2024-06-19 07:33:33+00:00

AI Summary

This systematic survey reviews Deepfake creation and detection methods. It categorizes Deepfake generation techniques and analyzes Deepfake detection approaches, highlighting challenges and future directions.

Abstract

Deepfake is a generative deep learning algorithm that creates or changes facial features in a very realistic way making it hard to differentiate the real from the fake features It can be used to make movies look better as well as to spread false information by imitating famous people In this paper many different ways to make a Deepfake are explained analyzed and separated categorically Using Deepfake datasets models are trained and tested for reliability through experiments Deepfakes are a type of facial manipulation that allow people to change their entire faces identities attributes and expressions The trends in the available Deepfake datasets are also discussed with a focus on how they have changed Using Deep learning a general Deepfake detection model is made Moreover the problems in making and detecting Deepfakes are also mentioned As a result of this survey it is expected that the development of new Deepfake based imaging tools will speed up in the future This survey gives indepth review of methods for manipulating images of face and various techniques to spot altered face images Four types of facial manipulation are specifically discussed which are attribute manipulation expression swap entire face synthesis and identity swap Across every manipulation category we yield information on manipulation techniques significant benchmarks for technical evaluation of counterfeit detection techniques available public databases and a summary of the outcomes of all such analyses From all of the topics in the survey we focus on the most recent development of Deepfake showing its advances and obstacles in detecting fake images

Key findings

The survey reveals that deep learning models, primarily CNNs, are predominantly used in Deepfake detection. While significant progress has been made, challenges remain in detecting increasingly sophisticated Deepfakes. The FF++ dataset is frequently utilized in research studies.

Approach

The paper conducts a systematic literature review, analyzing existing Deepfake creation and detection methods. It categorizes different Deepfake techniques and examines various detection approaches based on spatial, frequency, and temporal analysis, using deep learning models like CNNs, RNNs, and others.

Datasets

CelebA, UADFV, Deepfake-TIMIT, DFDC Preview, FaceForensics++, Celeb-DF, DeeperForensics-1.0, Wild Deepfake, OpenForensics, Diverse Fake Face Dataset (DFFD), Gender Balanced Deepfake Dataset (GBDF), DFDC, MICC (F220, F2000, F600), WWD, VISION, FaceFornesics (FF), etc.

Model(s)

Meso-4, MesoInception-4, CNN, LSTM, VGG-19, Capsule networks, SqueezeNet, ShallowNetV3, ResNetV2, Xception, FDFtNet, Xception network, SVM, Bayes classifier, VGG-16, ResNet, DNN, Logistic Regression, MLP, CLRNet, ResNet-18, VGG, ResNet-101, MDS network, Siamese network, Restricted Boltzmann Machine (RBM), EfficientNet, ResNext, AutoGAN, Xception-Net, VGG16, Autoencoder (AE), DeepRhythm, etc.

Author countries

Pakistan, UK, South Korea

← Previous