Recent Advances in Deepfake Detection: A Survey
Muhammad Haseeb Akbar, College of Aeronautical Engineering, National University of Sciences and Technology, Risalpur, Pakistan.
Fatima Khalid, Department of Computer Sciences, University of Engineering and Technology, Taxila, Pakistan.
Ammar Saleem, College of Aeronautical Engineering, National University of Sciences and Technology, Risalpur, Pakistan.
Nayyar Aafaq, Department of Computer Sciences, University of Engineering and Technology, Taxila, Pakistan.
Corresponding Author:
Fatima Khalid (fatimakhalid132@gmail.com)
Abstract:
The emergence of visual deepfakes, generated by advanced deep learning algorithms, has gained considerable attention in recent times. By utilizing deepfakes technology, facial manipulation can be achieved with remarkable precision, resulting in highly realistic alterations. Unfortunately, these videos have primarily targeted high-profile individuals such as politicians and celebrities and have been disseminated widely online. This has led to serious concerns about reputational damage and the manipulation of public opinion, which poses a significant threat to social stability. While the technology itself is neither good nor bad, it has frequently been utilized for nefarious purposes. Consequently, numerous studies have been undertaken to mitigate the negative impacts of deepfakes on society. These efforts have included the development of large-scale standards and the creation of detection techniques. In this paper, we aim to provide an overview of the current state of deepfakes image and video detection research, along with existing datasets. We also outline the challenges that must be considered to advance the field of deepfakes detection and identify potential future research directions. Although existing detection techniques have demonstrated some efficacy, they remain insufficient for real-world scenarios, and future research should focus on developing more resilient and generalizable detection methods.
Keywords:
Artificial Intelligence; Deepfakes; Deep Learning; Face Manipulation; GANs; Forgery Detection