Artificial Fingerprinting for Generative Models: Rooting Deepfake Attribution in Training Data

• Published in: Proceedings / IEEE International Conference on Computer Vision pp. 14428 - 14437
• Main Authors: Yu, Ning, Skripniuk, Vladislav, Abdelnabi, Sahar, Fritz, Mario
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.01.2021
• Subjects: accountability; and ethics in vision; Computational modeling; Data models; Fairness; Fingerprint recognition; Image and video manipulation detection and integrity methods; Technological innovation; Training data; transparency; Visualization; Weapons

Photorealistic image generation has reached a new level of quality due to the breakthroughs of generative adversarial networks (GANs). Yet, the dark side of such deepfakes, the malicious use of generated media, raises concerns about visual misinformation. While existing research work on deepfake det...