Deepfake Videos Research Articles

To Tell or Not to Tell: The Effects of Disclosing Deepfake Video on US and Indian Consumers’ Purchase Intention

Marketing communicators are utilizing artificial intelligence–driven technology to create seemingly real communications featuring synthetic humans. This research aims to determine the consequences of disclosing the use of synthetic, or “deepfake,” video avatars in marketing. Based on schema congruity theory and the persuasion knowledge model, this research investigates whether disclosing an avatar as synthetic impacts consumers’ perceptions of source credibility and subsequent purchase intention. It also examines the role of social cynicism on consumers’ perceptions of source credibility and purchase intention. This study employs a posttest only, control group experimental design (n = 318) with participants from the United States (low social cynicism) and India (high social cynicism). Using partial least squares structural equation modeling analysis, findings indicate that disclosure of a synthetic avatar negatively impacted purchase intention due to reduced perceptions of source trustworthiness. Results show that for women and higher-income individuals, disclosure has a greater negative effect on purchase intention. However, there was no significant support for social cynicism moderating the mediation effects of the three dimensions of source credibility on purchase intention. Discussion includes implications for researchers, practitioners, and policymakers interested in synthetic human avatar use in marketing communications.

A novel model compression method based on joint distillation for deepfake video detection

In recent years, deepfake videos have been abused to create fake news, which threaten the integrity of digital videos. Although existing detection methods leveraged cumbersome neural networks to achieve promising detection performance, they cannot be deployed in resource-constrained scenarios. To overcome this limitation, we propose a novel model compression framework based on joint distillation for deepfake detection, which includes a pre-training stage and a knowledge transfer stage. In the pre-training stage, a teacher network is trained with sufficient labeled samples. Then, in the knowledge transfer stage, a lightweight student network is constructed by considering dimension alignment. To transfer forensics knowledge comprehensively, a joint distillation loss is designed, including cross-entropy loss, knowledge distillation loss, and gradient-guided feature distillation loss. For feature distillation, feature maps from both shallow and deep layers are utilized to calculate channel-wise mean square error weighted by gradient information for transferring knowledge of forensics features adaptively. Besides, a decayed teaching strategy is constructed to adjust the importance of feature distillation, which aims at mitigating the risk of negative transfer. Extensive experiments show that student networks obtained by our model compression method can achieve competitive detection performance and outstanding efficiency by distinctly reducing computational costs compared against state-of-the-art methods.

Survey On “Deepvision’s Human Eye Blink Pattern Analysis for Deepfake Detection”

Deepfake technology has rapidly evolved, enabling the creation of highly convincing synthetic media, including images, videos, and audio recordings. These maliciously manipulated digital artifacts can have severe consequences for privacy, security, and trust in digital content. Therefore, the development of effective deepfake detection methods is of paramount importance. This paper proposes a novel approach for deepfake detection based on the analysis of human eye blinking patterns. The human eye blink is a unique and subtle behavioral trait that is challenging to replicate accurately in deepfake videos. We leverage this inherent biometric feature to identify anomalies in video content and distinguish between genuine and synthetic videos. Key Words: Eye Blink to Speech, Machine Learning.

Detecting Deepfake Videos Using Spatiotemporal Trident Network

The widespread dissemination of Deepfake in social networks has posed serious security risks, thus necessitating the development of an effective Deepfake detection technique. Currently, video-based detectors have not been explored as extensively as image-based detectors. Most existing video-based methods only consider temporal features without combining spatial features, and do not mine deeper-level subtle forgeries, resulting in limited detection performance. In this paper, a novel spatiotemporal trident network (STN) is proposed to detect both spatial and temporal inconsistencies of Deepfake videos. Since the large amount of redundant information in Deepfake video frames, we introduce convolutional block attention module (CBAM) on the basis of the I3D network and optimize the structure to make the network better focus on the meaningful information of the input video. Aiming at the defects in the deeper-level subtle forgeries, we designed three feature extraction modules (FEMs) of RGB, optical flow, and noise to further extract deeper video frame information. Extensive experiments on several well-known datasets demonstrate that our method has promising performance, surpassing several state-of-the-art Deepfake video detection methods.

Proposed DeepFake Detection Method Using Multiwavelet Transform

Videos made by artificial intelligence (A.I.) seem real, but they are not. When making DeepFake videos, face-swapping methods are frequently employed. The misuse of technology when using fakes, even though it was fun at first, these videos were somewhat recognizable to human eyes. However, as machine learning advanced, it became simpler to produce profound fake videos. It's practically impossible to tell it apart from actual videos now. Using GANs (Generative Adversarial Networks) and other deep learning techniques, DeepFake videos are output technology that may mislead people into thinking something is real when it is not. This study used a MultiWavelet transform to analyze the type of edge and its sharpness to develop a blur inconsistency detecting system. With this capability, it can assess whether or not the facial area is obscured in the video. As a result, it will detect fake videos. This paper reviews DeepFake detection techniques and discusses how they might be combined or altered to get more accurate results. A detection rate of more than 93.5% was obtained, which is quite successful.

MSVT: Multiple Spatiotemporal Views Transformer for DeepFake Video Detection

Recently, DeepFake videos have developed rapidly, causing new security issues in society. Due to the rough spatiotemporal view, existing video-based detection methods struggle to capture fine-grained spatiotemporal information, resulting in limited generalization ability. In addition, although the transformer has achieved great success in the past few years, the application of transformer on deepfake video detection still needs to be studied. To solve this problem, in this paper, we propose a novel Multiple Spatiotemporal Views Transformer (MSVT) with Local Spatiotemporal View (LSV) and Global Spatiotemporal View (GSV), to mine more detailed spatiotemporal information. Firstly, for establishing the LSV, different from existing works that sparsely sample a single frame to build the input sequence, we employ the local-consecutive temporal view to capture vital dynamic inconsistency. Furthermore, the extracted frame features within each group are fed to the temporal transformer followed by the feature fusion module, to generate group-level spatiotemporal features. Then, we further establish Global Spatiotemporal View (GSV) by feeding all the frame features within the whole video to the temporal transformer followed by the feature fusion module. Finally, we propose a novel global-local transformer (GLT) to effectively integrate these multi-level features for mining more subtle and comprehensive features. Extensive experiments on six large datasets demonstrate that our MSVT outperforms state-of-the-art detection methods.

Efficiency Evaluation of Popular Deepfake Methods Using Convolution Neural Network

Many deepfake techniques in the early years are spread to create successful deepfake videos (i.e., Face Swap, Deep Fake, etc.). These methods enable anyone to manipulate faces in videos, which can negatively impact society. One way to reduce this problem is the deepfake detection. It has become such a hot topic and the most crucial task in recent years. This paper proposes a deep learning model to detect and evaluate deepfake video methods using convolutional neural networks. The model is evaluated on the FaceForensics++ video dataset that contains four different deepfake ways (deepfake, face 2 face, face swap, and neuraltexture), and it achieved 0.96 accuracy on the deepfake method, 0.95 accuracy on face 2 face approach, 0.94 precision on face swap method and 0.76 accuracy on neuraltexture method.

DeepFake Detection with Remote Heart Rate Estimation Using 3D Central Difference Convolution Attention Network

Objective: As GAN-based deepfakes have become increasingly mature and realistic, the demand for effective deepfake detectors has become essential. We are inspired by the fact that normal pulse rhythms present in real-face video can be decreased or even completely interrupted in a deepfake video; thus, we have introduced a new deepfake detection approach based on remote heart rate estimation using the 3D Central Difference Convolution Attention Network (CDCAN). Methods: Our proposed fake detector is mainly composed of a 3D CDCAN with an inverse attention mechanism and LSTM architecture. It utilizes 3D central difference convolution to enhance the spatiotemporal representation, which can capture rich physiological-related temporal context by gathering the time difference information. The soft attention mechanism is to focus on the skin region of interest, while the inverse attention mechanism is to further denoise rPPG signals. Results: The performance of our approach is evaluated on the two latest Celeb-DF and DFDC datasets, for which the experiment results show that our proposed approach achieves an accuracy of 99.5% and 97.4%, respectively. Conclusion: Our approach outperforms the state-of-art methods and proves the effectiveness of our DeepFake detector.

Deepfake Satire and the Possibilities of Synthetic Media

This article explores the rise of deepfake satire as one of the most vibrant subgenres of experimentation within the expanding field of AI art. A combination of “deep learning” and the word “fake,” deepfake videos depict people doing or saying things that they never did or said. While deepfakes have been used to deceive and harm individuals as well as a mass audience, they have also been used toward alternative ends. Deepfake satire offers artful social critique, interrogating the individuals and institutions it portrays as much as the technology used to create it and the platforms on which it circulates. Videos range from snarky jabs at entertainment industry personalities to hard-hitting takedowns of tech entrepreneurs and authoritarian leaders. This speculative form of social imagining makes threatening and seemingly untouchable figures appear weak, vulnerable, and exposed, which can help envision a more equitable and just future. Close examination of specific projects within a broader media ecology leads to a nuanced understanding of both the menace and progressive possibilities of synthetic media—beyond simple “utopia”/“dystopia” binaries—in addition to expanding the grammar of computational art.

Deepfake Technology: A Criminological Literature Review

Cybercrimes are rapidly evolving with the advancement of technology and reaching new dimensions. Deepfake technology, which is the subject of this article, has also become a part of cybercrime. Deepfake technology, which has similar features to the CGI technology we have encountered in the cinema, is a technology that allows the image or voice of a person - or both - to be transferred to another image or video file in a hyper-realistic way and first came to the agenda in 2017. Although this technology has many benevolent uses, its malicious use is more prominent. This study aims to compile the criminological studies and current events in the literature on Deepfake for researchers who will study this field in the future. Within the scope of this purpose, this study examines the nature of Deepfake technology, the methods that can be used to detect Deepfake videos, the criminal dimensions of Deepfake technology and the violations it causes, the characteristics of victims and criminals, and the measures/senate resolutions taken by countries regarding the negative use of Deepfake technology.

Adjusting news accuracy perceptions after deepfakes exposure: Evidence from a non-Western context

Empirical studies examining if and how deepfakes would evoke news skepticism and discredit social trust in societies with low levels of digital literacy are limited. Recognizing the gap in the literature, this study conducts a survey experiment in India. We test whether individuals adjust their perceptions of news accuracy after being exposed to a deepfake about the subject of the news. Furthermore, we examine if cognitive ability will play a critical role in distinguishing how participants readjust their perceptions of news accuracy. In line with the propositions, we find that those exposed to a malicious deepfake video are primed to lower their initial positive news accuracy perceptions (featuring the subject) than those exposed to an informational video (featuring the subject). We also observe that the adjustments of news accuracy perceptions after being exposed to a deepfake are not dependent on the cognitive ability of social media news users. Finally, we discuss the theoretical and practical implications of our work.

All that glitters is not gold: a study of tourists' visit intention by watching deepfake destination videos

PurposeThis study aims to examine the tourists' visit intention by watching deepfake destination videos, using Information Manipulation and Media Richness Theory.Design/methodology/approachThis study conducted a primary survey utilizing a structured questionnaire. In total, 1,360 tourists were surveyed, and quantitative data analysis was done using PLS-SEM.FindingsThe results indicate that the factors that affect the tourists' visit intention after watching deepfake videos include information manipulation tactics, trust and media richness. This study also found that perceived deception and cognitive load do not influence the tourists' visit intention.Originality/valueThe originality/salience of this study lies in the fact that this is possibly among the first to combine the Media Richness Theory and Information Manipulation for understanding tourists' visit intention and post-viewing deepfake destination videos.

Providing detection strategies to improve human detection of deepfakes: An experimental study

Deepfake videos are becoming more pervasive. In this preregistered online experiment, participants (N = 454, Mage = 37.19, SDage = 13.25, males = 57.5%) categorize a series of 20 videos as either real or deepfake. All participants saw 10 real and 10 deepfake videos. Participants were randomly assigned to receive a list of strategies for detecting deepfakes based on visual cues (e.g., looking for common artifacts such as skin smoothness) or to act as a control group. Participants were also asked how confident they were that they categorized each video correctly (per video confidence) and to estimate how many videos they correctly categorized out of 20 (overall confidence). The sample performed above chance on the detection activity, correctly categorizing 60.70% of videos on average (SD = 13.00). The detection strategies intervention did not impact detection accuracy or detection confidence, with the intervention and control groups performing similarly on the detection activity and showing similar levels of confidence. Inconsistent with previous research, the study did not find that participants had a bias toward categorizing videos as real. Participants overestimated their ability to detect deepfakes at the individual video level. However, they tended to underestimate their abilities on the overall confidence question.

The Effect of Media Literacy on Misinformation and Deep Fake Video Detection

Misinformation has become a major problem in virtually all societies. The emergence of deepfake AI technology has only exacerbated issues around misinformation and disinformation. One way to try to limit the negative effects of fake news and false information in a society is media literacy. This study focused on the effect of media literacy on detecting fake news in both video and text messages. Based on Inoculation Theory, three hypotheses were suggested and a quasi-experimental research design was conducted. The participants included 204 highly educated Egyptians who were divided into three groups. One group was given a powerpoint lecture on the meaning of misinformation, its consequences, how to verify information received via social media, and examples of fake news. The second group was given a similar presentation but instead of talking about misinformation in general the presentation focused on deepfake AI technology. The third group was the control and were not given any media literacy information. Two types of fake messages were evaluated by each group; a Facebook post with text and an image and a deepfake video. The results of the study suggest the effectiveness of media literacy lectures in raising the rate of detection of fake news whether in the form of text or video. Results also indicated a decreased probability of sharing the fake messages. The study recommends designing media literacy lectures including both information on misinformation in general and specific information on the new technology creating fake news like deepfake.

Head Pose Estimation Patterns as Deepfake Detectors

The capacity to create ”fake” videos has recently raised concerns about the reliability of multimedia content. Identifying between true and false information is a critical step toward resolving this problem. On this issue, several algorithms utilizing deep learning and facial landmarks have yielded intriguing results. Facial landmarks are traits that are solely tied to the subject’s head posture. Based on this observation, we study how Head Pose Estimation (HPE) patterns may be utilized to detect deepfakes in this work. The HPE patterns studied are based on FSA-Net, SynergyNet, and WSM, which are among the most performant approaches on the state of the art. Finally, using a machine learning technique based on K-Nearest Neighbor and Dynamic Time Warping, their temporal patterns are categorized as authentic or false. We also offer a set of experiments for examining the feasibility of using deep learning techniques on such patterns. The findings reveal that the ability to recognize a deepfake video utilizing an HPE pattern is dependent on the HPE methodology. On the contrary, performance is less dependent on the performance of the utilized HPE technique. Experiments are carried out on the FaceForensics++ dataset, that presents both identity swap and expression swap examples. The findings show that FSA-Net is an effective feature extraction method for determining whether a pattern belongs to a deepfake or not. The approach is also robust in comparison to deepfake videos created using various methods or for different goals. In mean the method obtain 86% of accuracy on the identity swap task and 86.5% of accuracy on the expression swap. These findings offer up various possibilities and future directions for solving the deepfake detection problem using specialized HPE approaches, which are also known to be fast and reliable.

MRE-Net: Multi-Rate Excitation Network for Deepfake Video Detection

The current social media is flooded with hyper realistic face-synthetic videos due to the explosion of DeepFake technology that has brought a serious impact on human society security, which calls for further exploring on deepfake video detection methods. Existing methods attempt to isolated capture spatial artifacts or extract the homogeneous temporal inconsistency to detect deepfake video, but little attention has been paid to the exploitation of dynamic spatial-temporal inconsistency. To mitigate this issue, in this paper, we propose a novel Multi-Rate Excitation Network (MRE-Net) to effectively excite dynamic spatial-temporal inconsistency from the perspective of multiple rates for deepfake video detection. The proposed MRE-Net is composed of two components: Bipartite Group Sampling (BGS) and multiple rate branches. The BGS draws the entire video into multiple bipartite groups with different rates to cover various face motion dynamic evolution. We further design multiple rate branches to capture both short-term and long-term spatial-temporal inconsistency from corresponding bipartite groups of BGS. Concretely, for the early stages of the multi-rate branches, Momentary Inconsistency Excitation (MIE) module is developed to encode the spatial artifacts and intra-group short-term temporal inconsistency. Meanwhile, for the last stages of the multi-rate branches, Longstanding Inconsistency Excitation (LIE) module is constructed to perceive inter-group long-term temporal dynamics. Extensive experiments and visualizations conducted on four popular datasets demonstrate the effectiveness of the proposed method against state-of-the-art deepfake detection methods.

Jointly Defending DeepFake Manipulation and Adversarial Attack Using Decoy Mechanism.

Highly realistic imaging and video synthesis have become possible and relatively simple tasks with the rapid growth of generative adversarial networks (GANs). GAN-related applications, such as DeepFake image and video manipulation and adversarial attacks, have been used to disrupt and confound the truth in images and videos over social media. DeepFake technology aims to synthesize high visual quality image content that can mislead the human vision system, while the adversarial perturbation attempts to mislead the deep neural networks to a wrong prediction. Defense strategy becomes difficult when adversarial perturbation and DeepFake are combined. This study examined a novel deceptive mechanism based on statistical hypothesis testing against DeepFake manipulation and adversarial attacks. First, a deceptive model based on two isolated sub-networks was designed to generate two-dimensional random variables with a specific distribution for detecting the DeepFake image and video. This research proposes a maximum likelihood loss for training the deceptive model with two isolated sub-networks. Afterward, a novel hypothesis was proposed for a testing scheme to detect the DeepFake video and images with a well-trained deceptive model. The comprehensive experiments demonstrated that the proposed decoy mechanism could be generalized to compressed and unseen manipulation methods for both DeepFake and attack detection.

Artifact magnification on deepfake videos increases human detection and subjective confidence

Artifact magnification on deepfake videos increases human detection and subjective confidence

Deep Learning Methods to Detect Image Falsification

The accessibility and advancement of digital image editing tools have enabled individuals to manipulate and create realistic images without a real basis, leading to novel forms of creative expression and new professions. However, this also raises concerns over the malicious use of these technologies in spreading disinformation and fabricated evidence. Deepfake videos, which are generated using deep learning techniques, have become a major concern due to their potential to spread false information and cause harm to individuals and society as a whole. Therefore, the development of accurate and efficient deepfake detection methods has become an urgent need. After a thorough review of deep learning-based approaches for detecting deepfake videos, the LRNet method was chosen as a basis for further research due to its high precision. The method is designed to analyze the temporal changes in a video and identify whether the video has been manipulated or not. Experiments were conducted using a publicly available dataset. The first step involved analyzing the impact of model parameters. A total of 135 combinations were analyzed by changing the block size, dropout rate, learning rate, and optimizer. Based on the results, the model’s performance was enhanced by reducing the initial dropout rate, decreasing the number of GRU hidden neurons, and adding additional linear and ReLU6 layers. Upon conducting and comparing the results, it becomes evident that the chosen and improved method achieves promising outcomes in deepfake detection. This demonstrates the effectiveness of the dual-stream RNNs and the calibration module in enhancing the accuracy of geometric feature detection over time. This approach offers a more reliable way to detect deepfake videos, which is essential for preventing their malicious use in various domains. The results show that our proposed improved method achieves higher accuracy in some scenarios and provides a valuable analysis on how detection models are affected by their parameters.

Deepfake Video Detection via Facial Action Dependencies Estimation

Deepfake video detection has drawn significant attention from researchers due to the security issues induced by deepfake videos. Unfortunately, most of the existing deepfake detection approaches have not competently modeled the natural structures and movements of human faces. In this paper, we formulate the deepfake video detection problem into a graph classification task, and propose a novel paradigm named Facial Action Dependencies Estimation (FADE) for deepfake video detection. We propose a Multi-Dependency Graph Module (MDGM) to capture abundant dependencies among facial action units, and extracts subtle clues in these dependencies. MDGM can be easily integrated into the existing frame-level detection schemes to provide significant performance gains. Extensive experiments demonstrate the superiority of our method against the state-of-the-art methods.