Generative Adversarial Networks Applied for Privacy Preservation in Biometric-Based Authentication and Identification
This addresses privacy concerns for users of biometric authentication systems, though it is an incremental application of existing GAN techniques to a specific domain.
The paper tackles the problem of privacy leakage in biometric authentication systems by proposing a GAN-based method that translates face images to a visually private domain (e.g., flowers or shoes) for training classifiers, resulting in robustness against attacks while maintaining utility.
Biometric-based authentication systems are getting broadly adopted in many areas. However, these systems do not allow participating users to influence the way their data is used. Furthermore, the data may leak and can be misused without the users' knowledge. In this paper, we propose a new authentication method that preserves the privacy of individuals and is based on a generative adversarial network (GAN). Concretely, we suggest using the GAN for translating images of faces to a visually private domain (e.g., flowers or shoes). Classifiers, which are used for authentication purposes, are then trained on the images from the visually private domain. Based on our experiments, the method is robust against attacks and still provides meaningful utility.