Top Research Papers on Deepfake Detection

DeepMark is proposed, a scalable and robust framework for detecting DeepFakes that imprints essential visual features of a video into DeepMark Meta (DMM) and uses it to detect DeepFake manipulations by comparing the extracted visual features with the ground truth in DMM.

A detection model using convolution neural network (CNN) for face detection and Recurrent neural network (RNN) for video classification is developed for detecting fake videos.

This study proposes a pluggable and efficient active model watermarking framework for Deepfake detection that leverages the universal convolutional structure in generative model decoders and introduces convolutional kernel normalization to seamlessly integrate watermark parameters with those of the generative model.

This paper aims to investigate deepfake challenges, and to detect deepfake videos by using eye blinking, using convolutional neural networks to classify the eye states and long short term memory for sequence learning and the eye aspect ratio was used to calculate the height and width of open and closed eyes and to detects the blinking intervals.

This paper proposes a binary classifier based on a 2-phase learning architecture for detecting DeepFake images and demonstrates 91% validation accuracy on a large, diverse dataset of sophisticated GAN-generated DeepFake images.

This study conducts a comparative analysis of three varied convolutional neural networks (CNNs) for deepfake image detection by comparing three major CNN architectures and highlights the strengths and weaknesses of each CNN architecture.

The goal of the research is to create a deep learning-based method for identifying deepfake videos that can differentiate between real and fake information by using deep learning techniques and a variety of datasets for training, which helps combat the proliferation of false visual media.

This article explores the creation and detection of deepfakes and provides an in-depth view as to how these architectures work and the current trends and advancements in this domain.

This research introduces a Multimodal Deepfake Detection system capable of identifying manipulated content by combining visual and auditory cues, which employs convolutional neural networks to analyse video frames and process audio spectrograms, providing a comprehensive approach to detecting deepfake content.

This paper proposes a holistic strategy employing Facenet_pytorch, MTCNN, and InceptionResnetV1 for robust deepfake detection, achieving significant strides in differentiating manipulated content from authentic media, contributing to the ethical deployment of deepfake detection technologies.

This work develops an open-source online platform, known as DeepFake-o-meter, that integrates state-of-the-art DeepFake detection methods and provide a convenient interface for the users.

A study of the perceptions, current procedures, and expectations of journalists regarding a system for detection of deepfake videos, which combines technical knowledge of media forensics and the study to design a system usable by, and useful for, journalists.

The forgery detection process of the proposed DeepFake-Adapter model can be regularized by generalizable high-level semantics from a pre-trained ViT and adapted by global and local low-level forgeries of deepfake data.

In a fake video, the face, expression or speech is replaced with an image of another's face, with a distinct speech or emotion with the help of the technology of deep learning.

Analytic thinking and political interest were positively associated with identifying deepfakes and negatively associated with the perceived accuracy of a fake news piece about a leaked video (whether or not the deepfake video itself was presented).

EfficientNet-B5 is used to pluck out the spatial options of those faces they are fed as a batch of input series into a two-way long- and short-term memory (BiLSTM) to extract temporal characteristics.

This work introduces an open-source and user-friendly online platform, DeepFake-O-Meter v2.0, that integrates state-of-the-art methods for detecting DeepFake images, videos, and audio and serves as an evaluation and benchmanrking platform for researchers in digital media forensics to compare the performance of multiple algorithms on the same input.

A large-scale empirical study of facial deepfake/detection models and a set of key findings are drawn that the detection methods have poor generalization on samples generated by different deepfake methods, and there is no significant correlation between anti-detection ability and visual quality of deepfake samples.

This paper proposes an approach for detecting deepfake videos using Resnext CNN and LSTM, which can help in preventing the spread of misinformation and safeguarding the authors' society.

A wild testbed is built by gathering texts from various human writings and deepfake texts generated by different LLMs to showcase the challenges of deep-fake text detection in a wild testbed and show empirical results on automatic detection meth-ods.

This paper uses large margin cosine loss function (LMCL) and online frequency masking augmentation to force the neural network to learn more robust feature embeddings and achieves the lowest equal error rate (EER) among all single-system submissions during the ASVspoof 2019 challenge.

This work presents an alternative approach to DeepFake detection that focuses on whether the identity in the suspect image/video is true, and presents a simple identity-based detection algorithm called the OuterFace, which achieves superior detection accuracy and generalizes well to different DeepFake methods, and is robust with respect to video degradation techniques.

Various techniques that are used by several researchers to detect Deepfake videos, based on Machine Learning, Support Vector Machine and Deep Learning based techniques such as Convolution Neural Network with or without LSTM are analyzed.

The development of DeepFakes generation and detection technology is reviewed, and the existing approaches and datasets are systematically summarized and scientifically classified in order to provide a reference for further research in the field ofDeepFakes detection.

Evaluated methods of deepfake detection are evaluated and how they can be combined or modified to get more accurate results are discussed.

The proposed approach results in accuracy scores comparable to and surpassing several SOTA(State-of-the-art) approaches on three benchmark datasets, while consuming considerably lesser computational overhead, and containing over 100x lesser trainable parameters which was achieved using the extraction and manipulation of geometrical features.

This work has proposed a novel technique for recognizing FaceSwap-based deepfakes using landmarks computed from the input videos by employing Dlib-library and demonstrates that SVM works well than ANN in classifying the manipulated samples due to its power to deal with over-fitted training data.

A solution based on a GAN discriminator to detect DeepFake videos, which can be used to manipulate public opinion during elections, commit fraud, and discredit or blackmail people is explored.

Although Deep fake detection is extremely difficult and still an unsolved problem, a Deepfake detection model trained only on the DFDC can generalize to real "in-the-wild" Deepfake videos, and such a model can be a valuable analysis tool when analyzing potentially Deepfaked videos.

This work analyzes several deep learning approaches in the context of deepfakes classification in high compression scenarios and demonstrates that a proposed approach based on metric learning can be very effective in performing such a classification.

A novel research problem called Detecting Sequential DeepFake Manipulation (Seq-DeepFake), where face images are manipulated sequentially with corresponding annotations of sequential facial manipulation vectors, which requires correctly predicting a sequential vector of facial manipulation operations.

It is argued that the promise of algorithmic detectability falls short and that the ethico-politics of deepfakes cannot be reduced solely to a framework of detection algorithms.

This paper focuses on detecting DeepFake videos under three distinct scenarios, which are (i) all manipulation detection, (ii) single manipulation Detection, and then (iii) cross manipulation detection used to test the veracity of the videos.

This document finds DEFAKEHOP as the best method in deepfake detection, which uses spatial dimensionality reduction and soft classification to obtain a more concise description of the face for each detection channel.

The findings show that the detection are generally domain-specific tasks, however that using Transfer Learning considerably improves the model performance parameters, whereas convolutional RNN gives sequence detection advantage.

This paper seeks to delve to the nooks and crannies of the subject matter to provide a vivid understanding of biometric Spoofing and Deepfake Detection and throws light on a few areas where there exists the need for more research.

This study aims to delve deeper into the different methods used for detecting deepfakes, and assess their effectiveness in identifying manipulated content, including the quality of source material and the complexity of the deepfake algorithm.

The eye blinking pattern in deepfaked videos is not formed as naturally as it should be due to the incapability of Generative Adversarial Networks, necessitating the development of automated methods for detecting deep fake videos.

The methodologies used for deepfake detection using computer vision and deep learning on currently available datasets are evaluated and the results are presented in a comprehensive manner.

It has been revealed that current detection methods are still insufficient to be applied in real scenes, and further research should pay more attention to the generalization and robustness.

The survey shows that future research should address the lack of large scale datasets in the wild, poor generalization of existing detection methods to unknown fake attacks, as well as interpretability of detection results.

The Face Deepfake Detection and Reconstruction Challenge was described and two different tasks were proposed to the participants: creating a Deepfake detector capable of working in an “in the wild” scenario; and creating a method capable of reconstructing original images from Deepfakes.

This work performs the first large scale test of PRNU-based deepfake detection methods on a variety of standard datasets and shows the impact of often neglected parameters of the face extraction stage on detection accuracy.

A key component of FaceGuard is a new deep-learning-based watermarking method, which is robust to normal image post-processing such as JPEG compression, Gaussian blurring, cropping, and resizing, but fragile to deepfake manipulation.

The 1M-Deepfakes Detection Challenge is designed to engage the research community in developing advanced methods for detecting and localizing deepfake manipulations within the large-scale high-realistic audio-visual dataset.

A simple and efficient pipeline for detecting fake environmental sounds based on the CLAP audio embedding is proposed and it is shown that using an audio embedding trained specifically on environmental audio is beneficial over a standard VGGish one as it provides a 10% increase in detection performance.

This talk will describe the most reliable deep learning-based approaches for detecting deepfakes, with a focus on those that enable domain generalization, with reference to realistic scenarios, such as the dissemination of manipulated images and videos on social networks.

The Two-Stream models, which concurrently process spatial and temporal features, demonstrate superior performance in identifying deepfake videos, thereby underscoring the significance of temporal dynamics in deepfake detection technology.

This talk will describe the most reliable deep learning-based approaches for detecting deepfakes, with a focus on those that enable domain generalization, with reference to realistic scenarios, such as the dissemination of manipulated images and videos on social networks.

This is the first time to employ automated machine learning to adaptively search a neural architecture for deepfake detection, and based on the explored search space, this method achieves competitive prediction accuracy compared to previous methods.