HiDF: A Human-Indistinguishable Deepfake Dataset

DeepFake detection with multi-scale convolution and vision transformer

The rapid advancement of digital content generation technologies has led to the proliferation of deepfake media, posing significant security and privacy risks on social networks. Conventional deepfake detection methods struggle with accuracy, computational efficiency, and adaptability to evolving fake image generation techniques. To address these challenges, we propose a novel deepfake detection framework that integrates an Adaptive deep convolutional generative adversarial network (Adaptive-DCGAN) with a lightweight convolutional neural network (Lite-CNN). The Adaptive-DCGAN generates high-quality synthetic images, enhancing the training dataset, while the Lite-CNN efficiently classifies real and fake images with minimal computational overhead. Experimental results demonstrate that the proposed model achieves 95% accuracy, significantly outperforming existing GAN-CNN-based detection systems, which only reach 53% accuracy. Additionally, the framework exhibits superior performance metrics, including 95% precision, recall, and F1-score, ensuring robust and reliable deepfake detection. The proposed system offers a scalable and efficient solution for securing digital media platforms against deepfake threats. Future work will focus on improving generalization across diverse datasets, incorporating advanced synthetic image generation techniques, and enabling real-time deepfake detection for large-scale social media applications.

Learning a deep dual-level network for robust DeepFake detection

This research looks at the potential effects of generative artificial intelligence AI on the country's media landscape. Given their pervasiveness, it aims to reveal how AI-powered technologies in media content creation, distribution, and personalisation contribute to the overall process of national progress. Using well-designed questionnaires, the study quantitatively collects data from media professionals, techies, and communication scholars in large cities throughout China. Using statistical tools such as structural equation modelling and regression analysis, one investigated the interplay between the rate of modernisation, the effects of national development, and AI-driven media innovation. Media indices of generative AI demonstrate a clear positive correlation with the effect of modernism and national development programs. As China strives to digitally change its communication infrastructure and increase its cultural influence, technological prowess, and media production, generative AI is playing an increasingly crucial role. This study shows that AI in media may lead to more dynamic stories, practical audience participation, and worldwide outreach, all thanks to modernist techniques. There is no part of this that does not contribute to the advancement of national development goals. The results provide policymakers, media outlets, and AI developers with valuable information for formulating strategies to integrate AI with sustainable development objectives. Via an experimental interaction between generative AI and national development perceived via a modernist lens, this study provides a framework for future research on new media technologies and national change. The discussion of the societal potential presented by AI may now begin.

Abstract—Deepfake technology, driven by generative adversarial networks (GANs), poses significant challenges in digital security, misinformation, and privacy. Detecting deepfakes in images and videos requires advanced deep learning models. This study explores deepfake detection using convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformer-based architectures like Vision Transformers (ViTs). We employ Meso4_DF deepfake detection pipeline that uses TensorFlow/Keras, PyTorch, OpenCV for processing, with Dlib, Scikit-Image, and NumPy for feature extraction, leveraging transfer learning for enhanced accuracy. Video analysis includes frame extraction and temporal feature learning using LSTMs and 3D CNNs. Experimental results demonstrate that deep learning-based methods achieve high accuracy in distinguishing real and manipulated media, offering a robust approach to deepfake detection. Keywords—Deepfake Detection,CNN,GAN,Meso4 model, Image Analysis,Video Analysis.

Through the fast development of generative AI, it is now possible to create very realistic -deepfake tangibles, such as images, video, and audio, threatening the digital trust, media integrity, and cybersecurity. The article shows one of the instances of Hybrid AI-Based Multimodal Deepfake Detection System, analysis of images, video, and audio are provided in a single fusion model, which is beneficial to the check of high accuracy and power of detection. The system denotes dedicated modules of analysis of all the modalities like metadata validation and structural anomaly in images, temporal and frame based analysis in videos and acoustic pattern recognition in audio. The result of these modules is recombined with weighted hybrid decision fusion technique, to provide a synthesized output of detection. This hybrid method is less prone to false positive compared to its single-model counterpart and also more precise as cross-modal evidence is used. It was designed in a modular framework using Streamlit and enabled it to scale because further developments were enabled to use improved deep learning models, such as Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Transformer-based networks. It can be determined that unified hybrid methods have a better protection against compression artifacts, adversarial manipulations and data variability than the existing single-modality and partially hybrid methods. The findings indicate that multimodal hybrid reasoning is quite helpful in the improvement of the detection performance in various media manipulation. The article offers a far more adaptable and scalable structure of establishing resilient systems of deepfake detector that could keep up with the developing code of generative code and more advanced altered media techniques.

Internet video streaming applications have been demanding more bandwidth and higher video quality, especially with the advent of virtual reality and augmented reality appli-cations. While adaptive strea ming protocols like MPEG-DASH (dynamic adaptive streaming over HTTP) allows video quality to be flexibly adapted, e.g., degraded when mobile network condition deteriorates, this is not an option if the application itself requires guaranteed 4K quality at all time. On the other hand, conventional end-to-end transmission control protocol (TCP) has been struggling in supporting 4K video delivery across long-distance Internet paths containing both fixed and mobile network segments with heterogeneous characteristics. In this paper, we present a novel and practically feasible system architecture named MVP (mobile edge virtualization with adaptive prefetching), which enables content providers to embed their content intelligence as a virtual network function into the mobile network operator's infrastructure edge. Based on this architecture, we present a context-aware adaptive video prefetching scheme in order to achieve quality of experience (QoE)-assured 4K video on demand (VoD) delivery across the global Internet. Through experiments based on a real LTE-A network infrastructure, we demonstrate that our proposed scheme is able to achieve QoE-assured 4K VoD streaming, especially when the video source is located remotely in the public Internet, in which case none of the state-of-the-art solutions is able to support such an objective at global Internet scale.

The rapid development and availability of Generative AI (GenAI) tools, such as ChatGPT, offer new possibilities for academia, yet they also raise concerns for academic integrity. Chapelle's (2024) seminal work on GenAI’s implications serves as an impetus for this study, which explores teaching academic integrity in the context of AI use. Chapelle frames GenAI as an ‘inflection point’, calling for a re-evaluation of pedagogical strategies to support students in understanding its technological and ethical implications. This article presents findings from a mixed-methods study at a university in Cyprus, where ESL undergraduates in academic writing classes engaged with ChatGPT. Participants with B2/C1 English proficiency from diverse cultural backgrounds were assessed through pre- and post-intervention evaluations and qualitative interviews. A semester-long intervention incorporated activities to enhance academic writing skills and comprehension of ethical AI use, focusing on critical thinking, proper citation, and policy adherence. Findings indicate that students demonstrated increased AI knowledge and awareness of academic integrity. This study contributes to the discourse on integrating GenAI in academia, providing insights into effective instructional practices and addressing ethical considerations in academic writing. It offers guidance for educators and institutions aiming to integrate GenAI responsibly, ensuring academic integrity remains central in second language education.

AI-synthesized face-swapping videos, commonly known as DeepFakes, is an emerging problem threatening the trustworthiness of online information. The need to develop and evaluate DeepFake detection algorithms calls for large-scale datasets. However, current DeepFake datasets suffer from low visual quality and do not resemble DeepFake videos circulated on the Internet. We present a new large-scale challenging DeepFake video dataset, Celeb-DF, which contains 5, 639 high-quality DeepFake videos of celebrities generated using an improved synthesis process. We conduct a comprehensive evaluation of DeepFake detection methods and datasets to demonstrate the escalated level of challenges posed by Celeb-DF. Then we introduce Landmark Breaker, the first dedicated method to disrupt facial landmark extraction, and apply it to the obstruction of the generation of DeepFake videos. The experiments are conducted on three state-of-the-art facial landmark extractors using our Celeb-DF dataset.

Detecting digital face manipulation in images and video has attracted extensive attention due to the potential risk to public trust. To counteract the malicious usage of such techniques, deep learning-based deepfake detection methods have been employed and have exhibited remarkable performance. However, the performance of such detectors is often assessed on related benchmarks that hardly reflect real-world situations. For example, the impact of various image and video processing operations and typical workflow distortions on detection accuracy has not been systematically measured. In this paper, a more reliable assessment framework is proposed to evaluate the performance of learning-based deepfake detectors in more realistic settings. To the best of our acknowledgment, it is the first systematic assessment approach for deepfake detectors that not only reports the general performance under real-world conditions but also quantitatively measures their robustness toward different processing operations. To demonstrate the effectiveness and usage of the framework, extensive experiments and detailed analysis of four popular deepfake detection methods are further presented in this paper. In addition, a stochastic degradation-based data augmentation method driven by realistic processing operations is designed, which significantly improves the robustness of deepfake detectors.

The advent of Generative Artificial Intelligence (Generative AI or GAI) marks a significant inflection point in AI development. Long viewed as the epitome of reasoning and logic, Generative AI incorporates programming rules that are normative. However, it also has a descriptive component based on its programmers' subjective preferences and any discrepancies in the underlying data. Generative AI generates both truth and falsehood, supports both ethical and unethical decisions, and is neither transparent nor accountable. These factors pose clear risks to optimal decision-making in complex health services such as health policy and health regulation. It is important to examine how Generative AI makes decisions both from a rational, normative perspective and from a descriptive point of view to ensure an ethical approach to Generative AI design, engineering, and use. The objective is to provide a rapid review that identifies and maps attributes reported in the literature that influence Generative AI decision-making in complex health services. This review provides a clear, reproducible methodology that is reported in accordance with a recognised framework and Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 standards adapted for a rapid review. Inclusion and exclusion criteria were developed, and a database search was undertaken within four search systems: ProQuest, Scopus, Web of Science, and Google Scholar. The results include articles published in 2023 and early 2024. A total of 1,550articles were identified. After removing duplicates, 1,532articles remained. Of these, 1,511 articles were excluded based on the selection criteria and a total of 21 articles were selected for analysis. Learning, understanding, and bias were the most frequently mentioned Generative AI attributes. Generative AI brings the promise of advanced automation, but carries significant risk. Learning and pattern recognition are helpful, but the lack of a moral compass, empathy, consideration for privacy, and a propensity for bias and hallucination are detrimental to good decision-making. The results suggest that there is, perhaps, more work to be done before Generative AI can be applied to complex health services.

Modern AI models, such as large language models, are usually trained once on a huge corpus of data, potentially fine-tuned for a specific task, and then deployed with fixed parameters. Their training is costly, slow, and gradual, requiring billions of repetitions. In stark contrast, animals continuously adapt to the ever-changing contingencies in their environments. This is particularly important for social species, where behavioral policies and reward outcomes may frequently change in interaction with peers. The underlying computational processes are often marked by rapid shifts in an animal's behaviour and rather sudden transitions in neuronal population activity. Such computational capacities are of growing importance for AI systems operating in the real world, like those guiding robots or autonomous vehicles, or for agentic AI interacting with humans online. Can AI learn from neuroscience? This Perspective explores this question, integrating the literature on continual and in-context learning in AI with the neuroscience of learning on behavioral tasks with shifting rules, reward probabilities, or outcomes. We will outline an agenda for how specifically insights from neuroscience may inform current developments in AI in this area, and - vice versa - what neuroscience may learn from AI, contributing to the evolving field of NeuroAI.

The rapid development of deepfake technology has led to the generation of a large amount of tampered video and image content, posing a major challenge to content authenticity verification. In particular, detecting deepfakes in image sequences (e.g., agricultural product packaging) is particularly difficult because the anomalies introduced by the tampering techniques are often subtle and temporally continuous. In this paper, we propose a new deepfake detection method based on time series, combining independent component analysis (FastICA) with anomaly detection techniques. We first apply FastICA to extract independent components from image sequences to identify anomalous visual patterns that are unique to deepfake tampering. In addition, we use an efficient anomaly detection algorithm, LSHiforest, to achieve scalable and accurate identification of suspicious sequences. Experimental results show that the proposed method can still detect deepfake content with high accuracy in challenging scenarios with complex temporal dynamics. Our work provides a promising solution for real‐time and large‐scale detection of deepfake content in dynamic media.

In the digital era, deepfakes have become a critical threat to social media platforms and online information systems, undermining public trust and enabling misinformation campaigns at unprecedented scales. The advancements of artificial intelligence have promoted the rapid development of deepfake technology which enables the manipulation of visual content in highly realistic ways. While this technology holds potential for creative applications, it poses serious risks to digital platforms, public trust, and authenticity to online information. This growing concern demands efficient deepfake detection mechanisms. However, many studies primarily focus on facial manipulations, overlooking nonfacial alterations thereby leaving critical vulnerabilities unaddressed. So, this article proposes a novel approach that combines custom autoencoder (CA) based watermarking with postquantum cryptographic technique to detect the facial and nonfacial deepfakes. The proposed system detects deepfakes by extracting facial region as region of interest (ROI) and nonfacial region as nonregion of interest (NROI) using RetinaFace. A <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text{SHA}_{256}$</tex-math> </inline-formula> hash, h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{1}$</tex-math> </inline-formula> is generated for NROI and encrypting it using <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i>th-degree truncated polynomial ring (NTRU). This encrypted h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{1}$</tex-math> </inline-formula> is embedded into ROI using CA to generate watermarked image. During extraction, NROI of watermarked image is reprocessed to calculate a new hash, h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{2}$</tex-math> </inline-formula>. The encrypted h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{1}$</tex-math> </inline-formula> is extracted from ROI and decrypted to retrieve h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{1}$</tex-math> </inline-formula>, which is then compared with h<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}_{2}$</tex-math> </inline-formula>. Any mismatch reveals potential deepfakes. Experimental validation shows that our method achieves a high peak signal-to-noise ratio of 80.29 dB and an structural similarity index measure of 0.90, indicating that the watermark is invisible. It achieves 96% detection accuracy on our test set. Unlike most current methods that focus only on the face, our approach also identifies tampering with the body or background. Additionally, by using the NTRU encryption algorithm, it is protected against future quantum computer attacks. This makes it a practical and secure solution for safeguarding real-world digital media.

Computer vision is an interdisciplinary scientific field that deals with how computers obtain, store, interpret and understand digital images or videos using artificial intelligence based on neural networks, machine learning and deep learning methodologies. They are used in countless applications such as image retrieval and classification, driving and transport monitoring, medical diagnostics and aerial monitoring. Written by a team of international experts, this edited book covers the state-of-the-art of advanced research in the fields of computer vision and recognition systems from fundamental concepts to methodologies and technologies and real world applications including object detection, biometrics, Deepfake detection, sentiment and emotion analysis, traffic enforcement camera monitoring, vehicle control and aerial remote sensing imagery. The book will be useful for industry and academic researchers, scientists and engineers in the fields of computer vision, machine vision, image processing and recognition, multimedia, AI, machine and deep learning, data science, biometrics, security, and signal processing. It will also make a great course reference for advanced students and lecturers in these fields of research.