NullFace: Training-Free Localized Face Anonymization
Han-Wei Kung, Tuomas Varanka, Terence Sim, Nicu Sebe
2025-03-12
Summary
This paper talks about NullFace, a tool that hides faces in photos to protect privacy while keeping other details like expressions or features intact, using AI without needing special training.
What's the problem?
Current face-blurring or pixelating methods ruin photo quality and can’t let users choose which parts of the face to hide.
What's the solution?
NullFace uses a ready-made AI image generator to turn real faces into fake ones by changing their digital ‘fingerprint’, and lets users pick specific face areas to anonymize.
Why it matters?
This helps protect privacy in photos shared online or used in research while keeping the images clear and usable, like in medical studies or news reports.
Abstract
Privacy concerns around ever increasing number of cameras are increasing in today's digital age. Although existing anonymization methods are able to obscure identity information, they often struggle to preserve the utility of the images. In this work, we introduce a training-free method for face anonymization that preserves key non-identity-related attributes. Our approach utilizes a pre-trained text-to-image diffusion model without requiring optimization or training. It begins by inverting the input image to recover its initial noise. The noise is then denoised through an identity-conditioned diffusion process, where modified identity embeddings ensure the anonymized face is distinct from the original identity. Our approach also supports localized anonymization, giving users control over which facial regions are anonymized or kept intact. Comprehensive evaluations against state-of-the-art methods show our approach excels in anonymization, attribute preservation, and image quality. Its flexibility, robustness, and practicality make it well-suited for real-world applications. Code and data can be found at https://github.com/hanweikung/nullface .