Adversarial Makeup

Adversarial makeup is a research area focused on protecting facial privacy by generating digitally altered images that fool face recognition systems while maintaining a natural appearance. Current research employs generative adversarial networks (GANs) and diffusion models, often incorporating techniques like makeup transfer from reference images and 3D-aware generation to improve realism and robustness against various recognition systems. This work addresses the growing concern of unauthorized facial recognition, aiming to develop effective and visually plausible privacy-preserving methods with high transferability across different face recognition models and APIs. The success of these methods has significant implications for safeguarding individual privacy in the age of widespread facial recognition technology.