Face Anonymization Made Simple
This addresses privacy concerns in face image processing by providing a simpler and more effective anonymization method, with potential applications in face swapping.
The paper tackles the problem of face anonymization by proposing a diffusion model approach that eliminates the need for facial landmarks or masks, achieving state-of-the-art performance in identity anonymization, facial attribute preservation, and image quality on two public benchmarks.
Current face anonymization techniques often depend on identity loss calculated by face recognition models, which can be inaccurate and unreliable. Additionally, many methods require supplementary data such as facial landmarks and masks to guide the synthesis process. In contrast, our approach uses diffusion models with only a reconstruction loss, eliminating the need for facial landmarks or masks while still producing images with intricate, fine-grained details. We validated our results on two public benchmarks through both quantitative and qualitative evaluations. Our model achieves state-of-the-art performance in three key areas: identity anonymization, facial attribute preservation, and image quality. Beyond its primary function of anonymization, our model can also perform face swapping tasks by incorporating an additional facial image as input, demonstrating its versatility and potential for diverse applications. Our code and models are available at https://github.com/hanweikung/face_anon_simple .