الصفحة 1
الصفحة 1
Multimedia Forensics
Media forensics has never been more relevant to societal life. Not only media content represents an ever-increasing share of the data traveling on the net and the preferred communications means for most users, it has also become integral part of most innovative applications in the digital information ecosystem that serves various sectors of society, from the entertainment, to journalism, to politics. Undoubtedly, the advances in deep learning and computational imaging contributed significantly to this outcome. The underlying technologies that drive this trend, however, also pose a profound challenge in establishing trust in what we see, hear, and read, and make media content the preferred target of malicious attacks.
Handbook of Digital Face Manipulation and Detection: From DeepFakes to Morphing Attacks
Provides the first comprehensive collection of studies dealing with the hot topic of digital face manipulation such as DeepFakes, Face Morphing, or Reenactment. It combines the research fields of biometrics and media forensics including contributions from academia and industry. Appealing to a broad readership, introductory chapters provide a comprehensive overview of the topic, which address readers wishing to gain a brief overview of the state-of-the-art. Subsequent chapters, which delve deeper into various research challenges, are oriented towards advanced readers. Moreover, the book provides a good starting point for young researchers as well as a reference guide pointing at further literature. Hence, the primary readership is academic institutions and industry currently involved in digital face manipulation and detection. The book could easily be used as a recommended text for courses in image processing, machine learning, media forensics, biometrics, and the general security area.
Deepfake detection = اكتشاف التزييف العميق
In the rapidly evolving era of artificial intelligence, addressing the escalating threats of deepfake technology becomes a necessity because of the increasing sophistication of AI algorithms in generating deceptive content, and since it threatens the integrity of information across diverse data. The main objective is to build a sophisticated AI-driven system to detect different types of deepfake in text, audio, and images. In English text deepfake detection, multiple pre-trained tokenizers have been used, but XLNET and BERT stand out with identifying objects outside the dataset with an accuracy of 0.9809 and both have been generalized & trained using LSTM. In Arabic text deepfake detection, Arabert has been trained using LSTM which led with an accuracy of 99.53% by generalizing the model. Both English and Arabic datasets have been generated to enhance the accuracy and effectiveness of the models. Audio deepfake detection has been generalized too, using Random Forest with an accuracy of 98.259%.
Deepfake detection
Recently, various techniques of manipulating the video content have become available to everyone – online, one can find free applications e.g., for face swapping in videos. Such universal accessibility carries a notable risk of flooding online content with false information, affecting not only the greats of this world, but also the whole societies, also the rapid progress in synthetic image generation and manipulation has now come to a point where it raises significant concerns for the implications towards society. It is therefore necessary to develop a verification tool that will help assess the authenticity of the videos posted on the internet. This project describes the approach of using artificial intelligence solutions to detect doctored videos.
Deepfake detection
The rise of large language models (LLMs) and the increasing sophistication of deepfake images have made detecting synthetic content a pressing challenge. Several approaches have been proposed to tackle this problem, including statistical analysis, and machine learning algorithms. In this project, A novel zero-shot approach is proposed that utilizes the power of LLMs to detect fake text. The pre-trained LLM is fine-tuned to enhance its ability to differentiate real and fake text. The approach uses the LLM to detect text by analyzing the log probabilities of the text. For detecting fake images, computer vision algorithms and neural networks are used to analyze facial features. The facial region is cropped and preprocessed and the neural network identifies patterns indicative of synthetic content.
Deepfake
The technology used to create such digital content has quickly become accessible to the masses, such as “DEEPFAKE.” Deep fakes refer to manipulated videos, or other digital representations produced by sophisticated artificial intelligence, that yields to synthesize a sequence of face images and voices of characters corresponding to their identities, such as voice tone, facial expression, with a good lip synchronization. Therefore, this study is about developing real-time video generation software, which generates a target video from a single input image. Several methods and algorithms have been applied to detect, analyze personalize facial expression, voice and natural head poses to present a life-like image instead of a low quality one.
Deep fake detection
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.
