Deep learning has been successfully applied to solve various complex problems ranging from big data analytics to computer vision and human-level control. Deep learning advances however have also been employed to create software that can cause threats to privacy, democracy and national security. One of those deep learning-powered applications recently emerged is "deepfake". Deepfake algorithms can create fake images and videos that humans cannot distinguish them from authentic ones. The proposal of technologies that can automatically detect and assess the integrity of digital visual media is therefore indispensable. This paper presents a survey of algorithms used to create deepfakes and, more importantly, methods proposed to detect deepfakes in the literature to date. We present extensive discussions on challenges, research trends and directions related to deepfake technologies. By reviewing the background of deepfakes and state-of-the-art deepfake detection methods, this study provides a comprehensive overview of deepfake techniques and facilitates the development of new and more robust methods to deal with the increasingly challenging deepfakes.

Deep Learning for Deepfakes Creation and Detection: A Survey

Over the last few decades, rapid progress in AI, machine learning, and deep learning has resulted in new techniques and various tools for manipulating multimedia. Though the technology has been mostly used in legitimate applications such as for entertainment and education, etc., malicious users have also exploited them for unlawful or nefarious purposes. For example, high-quality and realistic fake videos, images, or audios have been created to spread misinformation and propaganda, foment political discord and hate, or even harass and blackmail people. The manipulated, high-quality and realistic videos have become known recently as Deepfake. Various approaches have since been described in the literature to deal with the problems raised by Deepfake. To provide an updated overview of the research works in Deepfake detection, we conduct a systematic literature review (SLR) in this paper, summarizing 112 relevant articles from 2018 to 2020 that presented a variety of methodologies. We analyze them by grouping them into four different categories: deep learning-based techniques, classical machine learning-based methods, statistical techniques, and blockchain-based techniques. We also evaluate the performance of the detection capability of the various methods with respect to different datasets and conclude that the deep learning-based methods outperform other methods in Deepfake detection.

/pdf/deepfake-detection-a-systematic-literature-review-wg98irxy.pdf

Deepfake Detection: A Systematic Literature Review

A review on flood management technologies related to image processing and machine learning

Creating deepfakes has rapidly become easier and more accessible due to advancements in hardware and computing. The harmful nature of deepfakes urges immediate action to improve detection of such doctored videos. In this work, we build a deep hybrid neural network model to detect deepfake videos. Using facial landmarks detection, we extract data pertaining to various facial attributes from the videos. This data is passed to a multilayer perceptron to learn differences in real and deepfake videos. Simultaneously, we use a convolutional neural network to extract features and train on the videos. We combine these two models to build a multi-input deepfake detector. A subset of the Deepfake Detection Challenge Dataset along with the Dessa Dataset is used to train the model. The proposed model provides good classification results with an accuracy of 84% and an AuC score of 0.87. 

Exposing deepfakes using a deep multilayer perceptron - convolutional neural network model

A comprehensive survey on digital video forensics: Taxonomy, challenges, and future directions

The spread of smartphones with high quality digital cameras in combination with easy access to a myriad of software apps for recording, editing and sharing videos and digital images in combination with deep learning AI platforms has spawned a new phenomenon of faking videos known as Deepfake. We design and implement a deep-fake detection model with mouth features (DFT-MF), using deep learning approach to detect Deepfake videos by isolating, analyzing and verifying lip/mouth movement. Experiments conducted against datasets that contain both fake and real videos showed favorable classification performance for DFT-MF model especially when compared with other work in this area.

Forensics and Analysis of Deepfake Videos

As Android is one of the most popular and widely used open-source mobile platforms, the security and privacy of Android apps are very critical, especially that over 6000 apps are added to the Google Play Store every day. This makes Android a prime target for malware. This paper proposes a modeling technique with experiments conducted using a dataset with about 10,000 benign and 10,000 malicious Android Application Packages (APK), in addition to other experiments that were conducted on the same dataset with a reduction in the number of benign files to be equal to 578 files. These files are analyzed using image classification techniques, where the whole APK file is converted into a grayscale image, and Convolutional Neural Networks (CNNs) with transfer-learning models are applied; to efficiently construct classification models for malware detection. Experiments have shown that the proposed technique has achieved favorable accuracy in the CNN model.

Malware Detection by Eating a Whole APK

Despite intensive efforts to prevent and control malicious human activities, they still pose grave risks and significant challenges in cyber-space. Malicious human activities are an evolving problem, especially with fast-growing technological advances. Sexual harassment or cyberbullying is considered an online malicious human activity that can easily affect legitimate users, governments, or other targets. The primary goal of this research is to propose an approach that could be utilized towards developing detection systems and enhance the classification of the different types of malicious human activities by using machine learning with different algorithms. Experiments showed that combining Term Frequency Inverse Document Frequency (TF-IDF) with machine learning achieved 81 % accuracy rate.

Sentiment Analysis-Based Sexual Harassment Detection Using Machine Learning Techniques

http://www.softxjournal.com/article/S2352711022000759/pdf

DFT-MF: Enhanced deepfake detection using mouth movement and transfer learning

Increasing water scarcity and rising demand throughout the Middle East and North Africa pose a major problem, and flood forecasting has been an open issue for a long time, attracting significant attention. Jordan seeks to use smart methods to solve the problem. Therefore, a real-world case study was conducted in Wadi al Wala for real-time rainfall forecasting and flood control, using 38 years of daily data from 13 rain gauge stations in the region. Different Machine Learning (ML) models were evaluated with various input information types to provide predictions in an almost real-time schedule. Preliminary tests showed that the decision tree (DT) and random forest (RF) techniques achieved the best generalized flood forecasting. In particular, the model was able to produce forecasts at any time, with the use of a mixture of meteorological parameters (relative humidity, air pressure, wet bulb temperature, and cloudiness), the precipitation at the forecasting point, and precipitation at the appropriate stations as input data, and the advanced ML model to be used with continuous data containing rainy and non-rainy cycles. Experiments showed the dominance of DT forecasts over those produced by the persistent model.

Mousa Tayseer Jafar

Papers

Forensics and Analysis of Deepfake Videos

Malware Detection by Eating a Whole APK

Sentiment Analysis-Based Sexual Harassment Detection Using Machine Learning Techniques

DFT-MF: Enhanced deepfake detection using mouth movement and transfer learning

Intelligent Methods for flood forecasting in Wadi al Wala, Jordan