DF-Net: The Digital Forensics Network for Image Forgery Detection
This addresses the societal threat of manipulated images spreading on online social networks, representing a domain-specific incremental improvement in digital forensics.
The paper tackles the problem of detecting manipulated images in online social networks by introducing DF-Net, a deep neural network for pixel-wise forgery detection, which outperforms state-of-the-art methods on four benchmark datasets and shows robustness against lossy image operations like resizing and compression.
The orchestrated manipulation of public opinion, particularly through manipulated images, often spread via online social networks (OSN), has become a serious threat to society. In this paper we introduce the Digital Forensics Net (DF-Net), a deep neural network for pixel-wise image forgery detection. The released model outperforms several state-of-the-art methods on four established benchmark datasets. Most notably, DF-Net's detection is robust against lossy image operations (e.g resizing, compression) as they are automatically performed by social networks.