Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper investigates the privacy vulnerabilities inherent in face-swapping anonymization techniques, revealing that significant identity leakage persists despite the visual similarity to donor faces. By employing a linear stochastic model to analyze the transformations applied by face-swappers on identity embeddings, the authors provide an interpretable framework for understanding and predicting this leakage. The findings emphasize the need for grounded privacy assessments, moving beyond observational metrics to establish formal, explainable privacy guarantees.
Identity leakage in face-swapping anonymization is not just a flaw—it's a predictable phenomenon that can be systematically analyzed and improved.
Face-swapping has emerged as a promising approach to facial privacy protection, replacing a target individual's appearance with that of a donor while preserving non-facial context. The resulting images visually resemble the donor, and face recognition systems tend to suppress the target's match scores -- ostensibly satisfying privacy requirements. Empirical evaluation across a range of face-swapping models, however, reveals that significant target identity leakage still occurs. This raises a deeper question: why does leakage occur, and can it be predicted? We propose a linear stochastic model that treats face-swappers as transformations on the space of identity embeddings, providing an interpretable account of the leakage mechanism. The model is fit to empirical observations and used to derive testable predictions. The aim is to ground privacy assessments in principled, interpretable analysis, thus making formal privacy guarantees explainable -- and perfectible -- rather than purely observational.