ROFI maintains 100% diagnostic sensitivity for eye diseases while anonymizing patient faces.
ROFI: a deep learning-based ophthalmic sign-preserving and reversible patient face anonymizer.
Yuan Tian et al. · NPJ digital medicine · 2025
Background
Patient face images provide a convenient mean for evaluating eye diseases, while also raising privacy concerns.
Results
diagnostic sensitivity
ROFI correctly found every case of eye disease
More results
Here, we introduce ROFI, a deep learning-based privacy protection framework for ophthalmology.
Using weakly supervised learning and neural identity translation, ROFI anonymizes facial features while retaining disease features (over 98% accuracy, κ > 0.90).
More results
ROFI works with AI systems, maintaining original diagnoses (κ > 0.80), and supports secure image reversal (over 98% similarity), enabling audits and long-term care.
“
Conclusion
These results show ROFI's effectiveness of protecting patient privacy in the digital medicine era.