Key points are not available for this paper at this time.
In recent years, the number of images and videos shared online increased and people have easy ways to access such content. “DeepFake” refers to any multimedia content created using deep learning technology in order to make it appear realistic. The creation of deepfake videos and images using deep learning techniques leads to very realistic “DeepFake” videos and images by changing the digital content of images and videos. Deepfake is widely recognized as one of artificial intelligence’s most dangerous uses. Deepfake makes it possible to place a person in a totally imaginary situation since it is used to imitate an activity that the person did not perform. Deepfakes have been becoming increasingly dangerous to democracy, society’s security and people’s privacy. The distribution of such deepfake content on various platforms urged the international community to revaluate the threat to social security posed by such content. It encouraged the researchers around the world to develop effective deepfake detection methods. In this paper we have discussed such approaches of deepfake detection in videos and images that are available in recent studies and have provided comparative review of research on deepfake detection algorithms. It also compares the different detection techniques and examines their limitations and advantages.
Kaushal et al. (Fri,) studied this question.