Year: 2024
Type: article
Abstract: With the rapid advancement of deep learning and generative modelling techniques, the creation of hyper-realistic synthetic media, known as deepfakes, has become a growing concern. These manipulated media assets are in various domains, including politics, entertainment, and security. As a result, the development of effective deepfake detection systems has gained significant attention. To identify Deepfakes, many researchers have developed various binary-classification-based detection techniques. This survey provides a comprehensive overview of state-of-the-art deepfake detection methods, their underlying principles, datasets used for training and evaluation, and the challenges faced in this evolving field. Then, alternative methods have been discussed in literature to address an issue raised by Deepfake. We analyse different techniques by groping them into four categories: Image-based DFDT, Audio-based DFDT, Video-based DFDT, and Multimodality based DFDT. Researchers in this area will learn from our study because it contains cutting-edge methods for detecting deep-fake videos and photos in social media. To benchmark and facilitate the advancement of deepfake detection systems, numerous datasets are examined, such as FaceForensics++, DeepFake Detection Challenge (DFDC), and Celeb-DF. Also discuss their characteristics, variations, and limitations, emphasizing the need for diverse and realistic datasets to ensure the model generalization.
Institution: Symbiosis International University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7218311915
