Deepfake Videos in the Wild: Analysis and Detection
This addresses the gap in evaluating deepfake detection methods for real-world applicability, which is crucial for combating misinformation and security threats, though it is incremental in focusing on dataset creation and evaluation rather than new detection algorithms.
The paper tackles the problem of limited understanding of real-world deepfake videos by collecting the largest dataset of 1,869 deepfake videos from YouTube and Bilibili, and finds that existing detection methods perform poorly on this dataset, indicating they are not ready for real-world deployment.
AI-manipulated videos, commonly known as deepfakes, are an emerging problem. Recently, researchers in academia and industry have contributed several (self-created) benchmark deepfake datasets, and deepfake detection algorithms. However, little effort has gone towards understanding deepfake videos in the wild, leading to a limited understanding of the real-world applicability of research contributions in this space. Even if detection schemes are shown to perform well on existing datasets, it is unclear how well the methods generalize to real-world deepfakes. To bridge this gap in knowledge, we make the following contributions: First, we collect and present the largest dataset of deepfake videos in the wild, containing 1,869 videos from YouTube and Bilibili, and extract over 4.8M frames of content. Second, we present a comprehensive analysis of the growth patterns, popularity, creators, manipulation strategies, and production methods of deepfake content in the real-world. Third, we systematically evaluate existing defenses using our new dataset, and observe that they are not ready for deployment in the real-world. Fourth, we explore the potential for transfer learning schemes and competition-winning techniques to improve defenses.