In this paper, we analyse the generalization ability of binary classifiers for the task of deepfake detection. We find that the stumbling block to their generalization is caused by the unexpected learned identity representation on images. Termed as the Implicit Identity Leakage, this phenomenon has been qualitatively and quantitatively verified among various DNNs. Furthermore, based on such understanding, we propose a simple yet effective method named the ID-unaware Deepfake Detection Model to reduce the influence of this phenomenon. Extensive experimental results demonstrate that our method outperforms the state-of-the-art in both in-dataset and cross-dataset evaluation. The code is available at https://github.com/megvii-research/CADDM.
Generative Adversarial Networks (GANs) have paved the path towards entirely new media generation capabilities at the forefront of image, video, and audio synthesis. However, they can also be misused and abused to fabricate elaborate lies, capable of stirring up the public debate. The threat posed by GANs has sparked the need to discern between genuine content and fabricated one. Previous studies have tackled this task by using classical machine learning techniques, such as k-nearest neighbours and eigenfaces, which unfortunately did not prove very effective. Subsequent methods have focused on leveraging on frequency decompositions, i.e., discrete cosine transform, wavelets, and wavelet packets, to preprocess the input features for classifiers. However, existing approaches only rely on isotropic transformations. We argue that, since GANs primarily utilize isotropic convolutions to generate their output, they leave clear traces, their fingerprint, in the coefficient distribution on sub-bands extracted by anisotropic transformations. We employ the fully separable wavelet transform and multiwavelets to obtain the anisotropic features to feed to standard CNN classifiers. Lastly, we find the fully separable transform capable of improving the state-of-the-art.
This paper describes the deepfake audio detection system submitted to the Audio Deep Synthesis Detection (ADD) Challenge Track 3.2 and gives an analysis of score fusion. The proposed system is a score-level fusion of several light convolutional neural network (LCNN) based models. Various front-ends are used as input features, including low-frequency short-time Fourier transform and Constant Q transform. Due to the complex noise and rich synthesis algorithms, it is difficult to obtain the desired performance using the training set directly. Online data augmentation methods effectively improve the robustness of fake audio detection systems. In particular, the reasons for the poor improvement of score fusion are explored through visualization of the score distributions and comparison with score distribution on another dataset. The overfitting of the model to the training set leads to extreme values of the scores and low correlation of the score distributions, which makes score fusion difficult. Fusion with partially fake audio detection system improves system performance further. The submission on track 3.2 obtained the weighted equal error rate (WEER) of 11.04\%, which is one of the best performing systems in the challenge.
Audio DeepFakes are utterances generated with the use of deep neural networks. They are highly misleading and pose a threat due to use in fake news, impersonation, or extortion. In this work, we focus on increasing accessibility to the audio DeepFake detection methods by providing SpecRNet, a neural network architecture characterized by a quick inference time and low computational requirements. Our benchmark shows that SpecRNet, requiring up to about 40% less time to process an audio sample, provides performance comparable to LCNN architecture - one of the best audio DeepFake detection models. Such a method can not only be used by online multimedia services to verify a large bulk of content uploaded daily but also, thanks to its low requirements, by average citizens to evaluate materials on their devices. In addition, we provide benchmarks in three unique settings that confirm the correctness of our model. They reflect scenarios of low-resource datasets, detection on short utterances and limited attacks benchmark in which we take a closer look at the influence of particular attacks on given architectures.
With the rise of AI-enabled Real-Time Deepfakes (RTDFs), the integrity of online video interactions has become a growing concern. RTDFs have now made it feasible to replace an imposter's face with their victim in live video interactions. Such advancement in deepfakes also coaxes detection to rise to the same standard. However, existing deepfake detection techniques are asynchronous and hence ill-suited for RTDFs. To bridge this gap, we propose a challenge-response approach that establishes authenticity in live settings. We focus on talking-head style video interaction and present a taxonomy of challenges that specifically target inherent limitations of RTDF generation pipelines. We evaluate representative examples from the taxonomy by collecting a unique dataset comprising eight challenges, which consistently and visibly degrades the quality of state-of-the-art deepfake generators. These results are corroborated both by humans and a new automated scoring function, leading to 88.6% and 80.1% AUC, respectively. The findings underscore the promising potential of challenge-response systems for explainable and scalable real-time deepfake detection in practical scenarios. We provide access to data and code at \url{https://github.com/mittalgovind/GOTCHA-Deepfakes}.
The increasing popularity of facial manipulation (Deepfakes) and synthetic face creation raises the need to develop robust forgery detection solutions. Crucially, most work in this domain assume that the Deepfakes in the test set come from the same Deepfake algorithms that were used for training the network. This is not how things work in practice. Instead, we consider the case where the network is trained on one Deepfake algorithm, and tested on Deepfakes generated by another algorithm. Typically, supervised techniques follow a pipeline of visual feature extraction from a deep backbone, followed by a binary classification head. Instead, our algorithm aggregates features extracted across all layers of one backbone network to detect a fake. We evaluate our approach on two domains of interest - Deepfake detection and Synthetic image detection, and find that we achieve SOTA results.
With the spread of DeepFake techniques, this technology has become quite accessible and good enough that there is concern about its malicious use. Faced with this problem, detecting forged faces is of utmost importance to ensure security and avoid socio-political problems, both on a global and private scale. This paper presents a solution for the detection of DeepFakes using convolution neural networks and a dataset developed for this purpose - Celeb-DF. The results show that, with an overall accuracy of 95% in the classification of these images, the proposed model is close to what exists in the state of the art with the possibility of adjustment for better results in the manipulation techniques that arise in the future.
The accelerated growth in synthetic visual media generation and manipulation has now reached the point of raising significant concerns and posing enormous intimidations towards society. There is an imperative need for automatic detection networks towards false digital content and avoid the spread of dangerous artificial information to contend with this threat. In this paper, we utilize and compare two kinds of handcrafted features(SIFT and HoG) and two kinds of deep features(Xception and CNN+RNN) for the deepfake detection task. We also check the performance of these features when there are mismatches between training sets and test sets. Evaluation is performed on the famous FaceForensics++ dataset, which contains four sub-datasets, Deepfakes, Face2Face, FaceSwap and NeuralTextures. The best results are from Xception, where the accuracy could surpass over 99\% when the training and test set are both from the same sub-dataset. In comparison, the results drop dramatically when the training set mismatches the test set. This phenomenon reveals the challenge of creating a universal deepfake detection system.
Deepfake refers to tailored and synthetically generated videos which are now prevalent and spreading on a large scale, threatening the trustworthiness of the information available online. While existing datasets contain different kinds of deepfakes which vary in their generation technique, they do not consider progression of deepfakes in a "phylogenetic" manner. It is possible that an existing deepfake face is swapped with another face. This process of face swapping can be performed multiple times and the resultant deepfake can be evolved to confuse the deepfake detection algorithms. Further, many databases do not provide the employed generative model as target labels. Model attribution helps in enhancing the explainability of the detection results by providing information on the generative model employed. In order to enable the research community to address these questions, this paper proposes DeePhy, a novel Deepfake Phylogeny dataset which consists of 5040 deepfake videos generated using three different generation techniques. There are 840 videos of one-time swapped deepfakes, 2520 videos of two-times swapped deepfakes and 1680 videos of three-times swapped deepfakes. With over 30 GBs in size, the database is prepared in over 1100 hours using 18 GPUs of 1,352 GB cumulative memory. We also present the benchmark on DeePhy dataset using six deepfake detection algorithms. The results highlight the need to evolve the research of model attribution of deepfakes and generalize the process over a variety of deepfake generation techniques. The database is available at: http://iab-rubric.org/deephy-database
Existing deepfake detectors face several challenges in achieving robustness and generalization. One of the primary reasons is their limited ability to extract relevant information from forgery videos, especially in the presence of various artifacts such as spatial, frequency, temporal, and landmark mismatches. Current detectors rely on pixel-level features that are easily affected by unknown disturbances or facial landmarks that do not provide sufficient information. Furthermore, most detectors cannot utilize information from multiple domains for detection, leading to limited effectiveness in identifying deepfake videos. To address these limitations, we propose a novel framework, namely Multimodal Graph Learning (MGL) that leverages information from multiple modalities using two GNNs and several multimodal fusion modules. At the frame level, we employ a bi-directional cross-modal transformer and an adaptive gating mechanism to combine the features from the spatial and frequency domains with the geometric-enhanced landmark features captured by a GNN. At the video level, we use a Graph Attention Network (GAT) to represent each frame in a video as a node in a graph and encode temporal information into the edges of the graph to extract temporal inconsistency between frames. Our proposed method aims to effectively identify and utilize distinguishing features for deepfake detection. We evaluate the effectiveness of our method through extensive experiments on widely-used benchmarks and demonstrate that our method outperforms the state-of-the-art detectors in terms of generalization ability and robustness against unknown disturbances.
Recently, Deepfake has drawn considerable public attention due to security and privacy concerns in social media digital forensics. As the wildly spreading Deepfake videos on the Internet become more realistic, traditional detection techniques have failed in distinguishing between real and fake. Most existing deep learning methods mainly focus on local features and relations within the face image using convolutional neural networks as a backbone. However, local features and relations are insufficient for model training to learn enough general information for Deepfake detection. Therefore, the existing Deepfake detection methods have reached a bottleneck to further improve the detection performance. To address this issue, we propose a deep convolutional Transformer to incorporate the decisive image features both locally and globally. Specifically, we apply convolutional pooling and re-attention to enrich the extracted features and enhance efficacy. Moreover, we employ the barely discussed image keyframes in model training for performance improvement and visualize the feature quantity gap between the key and normal image frames caused by video compression. We finally illustrate the transferability with extensive experiments on several Deepfake benchmark datasets. The proposed solution consistently outperforms several state-of-the-art baselines on both within- and cross-dataset experiments.
Human motion transfer refers to synthesizing photo-realistic and temporally coherent videos that enable one person to imitate the motion of others. However, current synthetic videos suffer from the temporal inconsistency in sequential frames that significantly degrades the video quality, yet is far from solved by existing methods in the pixel domain. Recently, some works on DeepFake detection try to distinguish the natural and synthetic images in the frequency domain because of the frequency insufficiency of image synthesizing methods. Nonetheless, there is no work to study the temporal inconsistency of synthetic videos from the aspects of the frequency-domain gap between natural and synthetic videos. In this paper, we propose to delve into the frequency space for temporally consistent human motion transfer. First of all, we make the first comprehensive analysis of natural and synthetic videos in the frequency domain to reveal the frequency gap in both the spatial dimension of individual frames and the temporal dimension of the video. To close the frequency gap between the natural and synthetic videos, we propose a novel Frequency-based human MOtion TRansfer framework, named FreMOTR, which can effectively mitigate the spatial artifacts and the temporal inconsistency of the synthesized videos. FreMOTR explores two novel frequency-based regularization modules: 1) the Frequency-domain Appearance Regularization (FAR) to improve the appearance of the person in individual frames and 2) Temporal Frequency Regularization (TFR) to guarantee the temporal consistency between adjacent frames. Finally, comprehensive experiments demonstrate that the FreMOTR not only yields superior performance in temporal consistency metrics but also improves the frame-level visual quality of synthetic videos. In particular, the temporal consistency metrics are improved by nearly 30% than the state-of-the-art model.
Fake content has grown at an incredible rate over the past few years. The spread of social media and online platforms makes their dissemination on a large scale increasingly accessible by malicious actors. In parallel, due to the growing diffusion of fake image generation methods, many Deep Learning-based detection techniques have been proposed. Most of those methods rely on extracting salient features from RGB images to detect through a binary classifier if the image is fake or real. In this paper, we proposed DepthFake, a study on how to improve classical RGB-based approaches with depth-maps. The depth information is extracted from RGB images with recent monocular depth estimation techniques. Here, we demonstrate the effective contribution of depth-maps to the deepfake detection task on robust pre-trained architectures. The proposed RGBD approach is in fact able to achieve an average improvement of 3.20% and up to 11.7% for some deepfake attacks with respect to standard RGB architectures over the FaceForensic++ dataset.
This paper presents our results and findings on the use of temporal images for deepfake detection. We modelled temporal relations that exist in the movement of 468 facial landmarks across frames of a given video as spatial relations by constructing an image (referred to as temporal image) using the pixel values at these facial landmarks. CNNs are capable of recognizing spatial relationships that exist between the pixels of a given image. 10 different ImageNet models were considered for the study.
In recent years, image and video manipulations with Deepfake have become a severe concern for security and society. Many detection models and datasets have been proposed to detect Deepfake data reliably. However, there is an increased concern that these models and training databases might be biased and, thus, cause Deepfake detectors to fail. In this work, we investigate factors causing biased detection in public Deepfake datasets by (a) creating large-scale demographic and non-demographic attribute annotations with 47 different attributes for five popular Deepfake datasets and (b) comprehensively analysing attributes resulting in AI-bias of three state-of-the-art Deepfake detection backbone models on these datasets. The analysis shows how various attributes influence a large variety of distinctive attributes (from over 65M labels) on the detection performance which includes demographic (age, gender, ethnicity) and non-demographic (hair, skin, accessories, etc.) attributes. The results examined datasets show limited diversity and, more importantly, show that the utilised Deepfake detection backbone models are strongly affected by investigated attributes making them not fair across attributes. The Deepfake detection backbone methods trained on such imbalanced/biased datasets result in incorrect detection results leading to generalisability, fairness, and security issues. Our findings and annotated datasets will guide future research to evaluate and mitigate bias in Deepfake detection techniques. The annotated datasets and the corresponding code are publicly available.
Deepfake media is becoming widespread nowadays because of the easily available tools and mobile apps which can generate realistic looking deepfake videos/images without requiring any technical knowledge. With further advances in this field of technology in the near future, the quantity and quality of deepfake media is also expected to flourish, while making deepfake media a likely new practical tool to spread mis/disinformation. Because of these concerns, the deepfake media detection tools are becoming a necessity. In this study, we propose a novel hybrid transformer network utilizing early feature fusion strategy for deepfake video detection. Our model employs two different CNN networks, i.e., (1) XceptionNet and (2) EfficientNet-B4 as feature extractors. We train both feature extractors along with the transformer in an end-to-end manner on FaceForensics++, DFDC benchmarks. Our model, while having relatively straightforward architecture, achieves comparable results to other more advanced state-of-the-art approaches when evaluated on FaceForensics++ and DFDC benchmarks. Besides this, we also propose novel face cut-out augmentations, as well as random cut-out augmentations. We show that the proposed augmentations improve the detection performance of our model and reduce overfitting. In addition to that, we show that our model is capable of learning from considerably small amount of data.
Face anti-spoofing (FAS) and face forgery detection play vital roles in securing face biometric systems from presentation attacks (PAs) and vicious digital manipulation (e.g., deepfakes). Despite promising performance upon large-scale data and powerful deep models, the generalization problem of existing approaches is still an open issue. Most of recent approaches focus on 1) unimodal visual appearance or physiological (i.e., remote photoplethysmography (rPPG)) cues; and 2) separated feature representation for FAS or face forgery detection. On one side, unimodal appearance and rPPG features are respectively vulnerable to high-fidelity face 3D mask and video replay attacks, inspiring us to design reliable multi-modal fusion mechanisms for generalized face attack detection. On the other side, there are rich common features across FAS and face forgery detection tasks (e.g., periodic rPPG rhythms and vanilla appearance for bonafides), providing solid evidence to design a joint FAS and face forgery detection system in a multi-task learning fashion. In this paper, we establish the first joint face spoofing and forgery detection benchmark using both visual appearance and physiological rPPG cues. To enhance the rPPG periodicity discrimination, we design a two-branch physiological network using both facial spatio-temporal rPPG signal map and its continuous wavelet transformed counterpart as inputs. To mitigate the modality bias and improve the fusion efficacy, we conduct a weighted batch and layer normalization for both appearance and rPPG features before multi-modal fusion. We find that the generalization capacities of both unimodal (appearance or rPPG) and multi-modal (appearance+rPPG) models can be obviously improved via joint training on these two tasks. We hope this new benchmark will facilitate the future research of both FAS and deepfake detection communities.
Recently, pioneer research works have proposed a large number of acoustic features (log power spectrogram, linear frequency cepstral coefficients, constant Q cepstral coefficients, etc.) for audio deepfake detection, obtaining good performance, and showing that different subbands have different contributions to audio deepfake detection. However, this lacks an explanation of the specific information in the subband, and these features also lose information such as phase. Inspired by the mechanism of synthetic speech, the fundamental frequency (F0) information is used to improve the quality of synthetic speech, while the F0 of synthetic speech is still too average, which differs significantly from that of real speech. It is expected that F0 can be used as important information to discriminate between bonafide and fake speech, while this information cannot be used directly due to the irregular distribution of F0. Insteadly, the frequency band containing most of F0 is selected as the input feature. Meanwhile, to make full use of the phase and full-band information, we also propose to use real and imaginary spectrogram features as complementary input features and model the disjoint subbands separately. Finally, the results of F0, real and imaginary spectrogram features are fused. Experimental results on the ASVspoof 2019 LA dataset show that our proposed system is very effective for the audio deepfake detection task, achieving an equivalent error rate (EER) of 0.43%, which surpasses almost all systems.
Deepfakes are the synthesized digital media in order to create ultra-realistic fake videos to trick the spectator. Deep generative algorithms, such as, Generative Adversarial Networks(GAN) are widely used to accomplish such tasks. This approach synthesizes pseudo-realistic contents that are very difficult to distinguish by traditional detection methods. In most cases, Convolutional Neural Network(CNN) based discriminators are being used for detecting such synthesized media. However, it emphasise primarily on the spatial attributes of individual video frames, thereby fail to learn the temporal information from their inter-frame relations. In this paper, we leveraged an optical flow based feature extraction approach to extract the temporal features, which are then fed to a hybrid model for classification. This hybrid model is based on the combination of CNN and recurrent neural network (RNN) architectures. The hybrid model provides effective performance on open source data-sets such as, DFDC, FF++ and Celeb-DF. This proposed method shows an accuracy of 66.26%, 91.21% and 79.49% in DFDC, FF++, and Celeb-DF respectively with a very reduced No of sample size of approx 100 samples(frames). This promises early detection of fake contents compared to existing modalities.
In the recent years, social media has grown to become a major source of information for many online users. This has given rise to the spread of misinformation through deepfakes. Deepfakes are videos or images that replace one persons face with another computer-generated face, often a more recognizable person in society. With the recent advances in technology, a person with little technological experience can generate these videos. This enables them to mimic a power figure in society, such as a president or celebrity, creating the potential danger of spreading misinformation and other nefarious uses of deepfakes. To combat this online threat, researchers have developed models that are designed to detect deepfakes. This study looks at various deepfake detection models that use deep learning algorithms to combat this looming threat. This survey focuses on providing a comprehensive overview of the current state of deepfake detection models and the unique approaches many researchers take to solving this problem. The benefits, limitations, and suggestions for future work will be thoroughly discussed throughout this paper.