Forgery Domain
The forgery domain focuses on detecting manipulated media, primarily deepfakes, which pose significant societal risks. Current research emphasizes developing robust and generalizable detection models, often employing deep learning architectures like transformers and incorporating techniques such as federated learning for privacy preservation and incremental learning to adapt to evolving forgery methods. Key advancements involve leveraging diverse data representations, including frequency domain analysis alongside traditional image analysis, and designing methods to improve model generalization across different forgery techniques and datasets. This field is crucial for mitigating the spread of misinformation and maintaining trust in digital media.
Papers
Unified Frequency-Assisted Transformer Framework for Detecting and Grounding Multi-Modal Manipulation
Huan Liu, Zichang Tan, Qiang Chen, Yunchao Wei, Yao Zhao, Jingdong Wang
DFIL: Deepfake Incremental Learning by Exploiting Domain-invariant Forgery Clues
Kun Pan, Yin Yifang, Yao Wei, Feng Lin, Zhongjie Ba, Zhenguang Liu, ZhiBo Wang, Lorenzo Cavallaro, Kui Ren