AuraMask: Aesthetic Anti-Facial Recognition Filters for User Acceptance
Researchers propose AuraMask, a pipeline for generating anti-facial recognition (AFR) image filters that are both effective against computer vision and aesthetically acceptable to users. The system produces 40 filters emulating popular Instagram one-click filters. In tests, AuraMask filters matched or exceeded prior methods in adversarial effectiveness against open-source facial recognition models. A controlled online user study (N=630) confirmed significantly higher user acceptance compared to previous AFR approaches. The pipeline is released to the community to encourage broader adoption of privacy-preserving image alterations that do not compromise self-presentation.
Key facts
- AuraMask is a pipeline for developing aesthetic anti-facial recognition (AFR) image filters.
- The filters are designed to be subtle to humans but blinding to computer vision.
- 40 aesthetic filters were produced that emulate popular Instagram one-click filters.
- AuraMask filters meet or exceed adversarial effectiveness of prior methods against open-source facial recognition models.
- A controlled online user study (N=630) showed significantly higher user acceptance than prior methods.
- The AFR pipeline is provided to the community.
- The research addresses the conflict between privacy and self-presentation in AFR filters.
- The study was published on arXiv (2605.12937).
Entities
Institutions
- arXiv