This Face Does Not Exist... But It Might Be Yours! Identity Leakage in Generative Models

• Published in: Proceedings / IEEE Workshop on Applications of Computer Vision pp. 1319 - 1327
• Main Authors: Tinsley, Patrick, Czajka, Adam, Flynn, Patrick
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.01.2021
• Subjects: Data models; Faces; Generative adversarial networks; Image resolution; Training; Training data

Generative adversarial networks (GANs) are able to generate high resolution photo-realistic images of objects that "do not exist." These synthetic images are rather difficult to detect as fake. However, the manner in which these generative models are trained hints at a potential for inform...