Refacing: reconstructing anonymized facial features using GANs

TL;DR

This paper demonstrates that current anonymization techniques for medical images, like face blurring and removal, can be partially reversed using GANs, raising concerns about data privacy and the effectiveness of anonymization methods.

Contribution

The study applies CycleGAN to reconstruct facial features from anonymized MR images, revealing vulnerabilities in existing anonymization techniques.

Findings

Face blurring is insufficient for privacy protection.

Face removal offers better anonymization but remains partially reversible.

GAN-based methods can reconstruct identifiable facial features.

Abstract

Anonymization of medical images is necessary for protecting the identity of the test subjects, and is therefore an essential step in data sharing. However, recent developments in deep learning may raise the bar on the amount of distortion that needs to be applied to guarantee anonymity. To test such possibilities, we have applied the novel CycleGAN unsupervised image-to-image translation framework on sagittal slices of T1 MR images, in order to reconstruct facial features from anonymized data. We applied the CycleGAN framework on both face-blurred and face-removed images. Our results show that face blurring may not provide adequate protection against malicious attempts at identifying the subjects, while face removal provides more robust anonymization, but is still partially reversible.